The Future of Last-Mile Delivery: a Scenario Thinking Approach

THE FUTURE OF LAST-MILE DELIVERY: A SCENARIO THINKING APPROACH

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Kolawole Samuel Ewedairo B.Sc. (Hons) Geography and Regional Planning, OOU, Nigeria Master of Urban and Regional Planning, University of Ibadan, Nigeria

School of Business IT and Logistics College of Business RMIT University

June 2019

DECLARATION

I certify that except where due acknowledgement has been made, the work is that of the author alone; the work has not been submitted previously, in whole or in part, to qualify for any other academic award; the content of the project is the result of work which has been carried out since the official commencement date of the approved research program; any editorial work, paid or unpaid, carried out by a third party is acknowledged; and, ethics procedures and guidelines have been followed.

I acknowledge the support I have received for my research through the provision of an Australian Government Research Training Program Scholarship.

Kolawole Ewedairo June 2019

ACKNOWLEDGEMENTS

To God who is the same yesterday, today and forever be the glory, honour and adoration.

First and foremost, my deepest appreciation goes to my senior supervisor, Professor Prem Chhetri. You are mentor, a friend and a brother. No words of gratitude would suffice for the precious guidance, continuous support, and attention to detail, motivation, inspiration, and patience that you provided throughout this PhD journey. I thank God for your life and that of your entire family. May God indeed bless you and keep you and your family. I would also like to express my appreciation to my associate supervisor Professor Jago Dodson for his support. You are a great scholar indeed. To Professor Lee Taewoo and Associate Professor Ferry Jie thank you for your contribution to the success of this journey. God still answers prayers. Professor Shams Rahman your contribution to the success of this journey is appreciated. I would also like to thank Professor Caroline Chan, former Head of School, School of Business IT & Logistics at RMIT, for her generous support. Similarly, gratitude is due to Associate Professor Victor Gekara, Dr Huhammad Abdulrahman, and Dr Sharoooz, for your support. To all other academics and friends in BITL, I say a big thank you for your support and encouragement. Thank you all.

To my wife Oluwafunmilayo, my children Oluwapamilerin, Oluwasemiloore and Oluwatemilorun, thank you for the encouragement to commence the journey as well as your understanding and support during the journey. Hi fives to each one of you, we have made it. To Muyiwa and Mosun Agunbiade, word is not enough to show my appreciation. Thank you to Anjali, Anya and Eveny for allowing me to be a part of your family where I belong forever. Professors Kayode Oyesiku and Dele Badejo, Kayode and Bola Adeyemi, Tunde and Molayo Adedotun, Laoye and Femi Sopade, Gbenga and Bola Tokun, Dare and Tokunbo Ogunmola, thank you all for your supports and prayers. Appreciation to Pastors Ibukun and Shalewa Williams, Pastors Samuel and Mary Olaniyan and the entire members of RCCG Jesus House Melbourne.

I thank both Wuraola Ajoke and Olusola Ewedairo, I will never forget their sacrifice. Continue to rest in the peace. Your effort has been crowned with the success of this journey.

To all who space will not allow me to mention here, thank you for your support. I will ever be grateful to all of you.

iii

TABLE OF CONTENTS

DECLARATION ...... i ACKNOWLEDGEMENTS ...... ii TABLE OF CONTENTS ...... iv LIST OF TABLES ...... ix LIST OF FIGURES ...... x LIST OF ABBREVIATIONS ...... xiv LIST OF PUBLICATIONS ...... xvi ABSTRACT ...... xvii CHAPTER ONE ...... 1 INTRODUCTION ...... 1 1.1 Introduction ...... 1 1.2 Problem Statement ...... 2 1.3 Research Rationale ...... 5 1.3.1 Increased Population Growth and Growing Pressure on Cities ...... 6 1.3.2 The Changing Built and Regulatory Environment under the Compact Cities Model ...... 9 1.3.3 A Mismatch between Consumer Demand for Goods and Transportation Infrastructure Supply ...... 10 1.3.4 Changing Consumer Demand and Distribution Methods ...... 11 1.4 Scope of Study ...... 16 1.5 Research Aim and Questions ...... 17 1.6 Research Methodology ...... 17 1.6.1 Study Area ...... 17 1.6.2 Methods...... 18 1.7 Thesis Structure ...... 19 1.8 Summary ...... 21 CHAPTER TWO ...... 23 UNDERSTANDING LAST MILE DELIVERY IMPEDANCE ...... 23 2.1 Introduction ...... 23 2.2 City Logistics ...... 23 2.2.1 Last Mile Delivery ...... 25 2.2.2 Characteristics of Last Mile Logistics ...... 27 2.3 Perspectives on Last Mile Delivery ...... 30

2.3.1 Supply Chain Perspective ...... 30 2.3.2 Transaction Cost Perspective ...... 31 2.3.3 Spatial Perspective ...... 33 2.3.4 Urban Planning Perspective ...... 34 2.4 Urban Form and Last Mile Logistics ...... 35 2.5 Last Mile Delivery Impedance ...... 40 2.6 Summary ...... 41 CHAPTER THREE ...... 43 CONCEPTUAL FRAMEWORK FOR BUILDING THE FUTURE OF LAST MILE DELIVERY...... 43 3.1 Introduction ...... 43 3.2 A Theoretical Framework for Building LMD Scenarios ...... 43 3.3 Theory of Constraints (TOC) ...... 47 3.3.1 The First Principle: The Logistics Paradigm ...... 47 3.3.2 The Second Principle: The Thinking Process (TP)...... 49 3.4 Scenario Thinking Method ...... 51 3.4.1 Scenario Thinking - Schools of Thought ...... 53 3.4.2 Methods and Techniques to Building Scenarios ...... 54 3.4.3 Time Horizon in Scenario Thinking ...... 58 3.5 Identifying Last Mile Delivery Constraints ...... 58 3.5.1 Population Density ...... 61 3.5.2 Activity Centre ...... 63 3.5.3 Tollways ...... 64 3.5.4 Speed limit ...... 66 3.5.5 Traffic lights...... 66 3.5.6 Landuse Zoning ...... 67 3.5.7 Traffic Count ...... 68 3.5.8 Number of Lanes...... 68 3.5.9 Tram Lanes ...... 69 3.5.10 Railway Boomgate or Level Crossing ...... 70 3.5.11 Bicycle Lanes ...... 72 3.5.12 Other Constraints ...... 72 3.6 Summary ...... 73 CHAPTER FOUR ...... 74 RESEARCH METHODOLOGY...... 74

4.1. Introduction ...... 74 4.2 The Approach ...... 74 4.3 Research Methodology ...... 77 4.3.1 Study Area ...... 78 4.3.2 Key Case Studies ...... 85 4.3.3 Scenario Thinking Framework ...... 88 4.3.4 Scenario Thinking Workshop ...... 89 4.4 Research Design ...... 93 4.4.1 Stage 1: Identification of Key Constraints ...... 95 4.4.2 Stage Two: Clustering and Defining of Dimensions ...... 105 4.4.3 Stage 3: Impact and Level of Uncertainty ...... 106 4.4.4 Stage 4: Scoping the Future Last Mile Delivery Scenarios ...... 108 4.4.5 Stage 5. Stakeholder Analysis...... 109 4.5 Ethical Considerations...... 111 4.6 Summary ...... 113 CHAPTER FIVE ...... 114 IDENTIFICATION AND ESTIMATION OF LAST MILE DELIVERY IMPEDANCE ... 114 5.1 Introduction ...... 114 5.2 Stage One: Identification and Estimation of Last Mile Delivery Impedance ...... 114 5.2.1 Task 1: Identification of Last Mile Delivery Constraints ...... 115 5.2.2 Task 2: Assessment and Standardisation of Spatial Constraint Data ...... 117 5.2.3 Task 3: Estimation and Mapping of the Last Mile Delivery Impedance Index ...... 119 5.3 Metropolitan Last Mile Delivery Wide Analysis ...... 126 5.3.1 Population Density ...... 130 5.3.2 Proximity to Activity Centres ...... 131 5.3.3 Tollways ...... 131 5.3.4 Landuse Zoning ...... 132 5.3.5 Speed limits ...... 132 5.3.6 Trams and Railway Boomgate ...... 133 5.4 Local Area Analysis ...... 135 5.5 Summary ...... 139

CHAPTER SIX ...... 140 BUILDING LAST MILE DELIVERY SCENARIOS ...... 140 6.1 Introduction ...... 140 6.2 Building Last Mile Delivery Scenarios ...... 140 6.3 Stage Two: Clustering of Last Mile Delivery Constraints ...... 141 6.3.1 Freight Infrastructure ...... 142 6.3.2 Landuse Intensity ...... 144 6.3.3 Infrastructure Supply ...... 146 6.3.4 Intersection Controls...... 148 6.3.5 Infrastructure Sharing ...... 149 6.3.6 Human Behaviour ...... 150 6.4 Stage Three: Generating Impact and Uncertainty Matrix ...... 151 6.5 Stage Four: Scoping and Developing LMD Scenarios ...... 152 6.5.1 Best/Worst Scenario...... 155 6.5.2 Best/Best Scenario ...... 156 6.5.3 Worst/Worst Scenario ...... 157 6.5.4 Worst/Best Scenario...... 158 6.6: Stage Five: Stakeholder Analysis - Power and Interest ...... 159 6.6.1 Quadrant 1 - High Power-Low Interest ...... 160 6.6.2 Quadrant 2 - High Power-High Interest ...... 162 6.6.3 Quadrant 3 - Low Power-Low Interest ...... 163 6.6.4 Quadrant 4 - Low Power-High Interest ...... 163 6.7 Summary ...... 164 CHAPTER SEVEN ...... 165 STRATEGIC FRAMEWORK FOR LAST MILE DELIVERY ...... 165 7.1 Introduction ...... 165 7.2 Development of a Strategic Framework for LMD ...... 165 7.2.1 Setting the Last Mile Delivery Goals...... 167 7.2.2 Desirable Distribution Objectives (DDO) ...... 168 7.3 Last Mile Delivery Strategies ...... 169 7.3.1 Landuse Zoning Strategy ...... 171 7.3.2 Last Mile Delivery Corridor Strategy ...... 173 7.3.3 Distribution Network Strategy ...... 174 7.3.4 Multi-Modal Use Strategy ...... 176 7.3.5 Stakeholder Engagement Strategy ...... 177

vii

7.4 Possible Actions (PA) ...... 179 7.4.1 PA 1: Regulating the Size of Vehicles and Delivery Time ...... 180 7.4.2 PA 2: Consolidation of load and Load Factor Efficiency ...... 181 7.4.3 PA 3: Urban Consolidation Centre (UCC) ...... 181 7.4.4 PA 4: Shared Economy Business and Information Technology Systems ...... 184 7.4.5 PA 5: Underground Loading and Unloading Area ...... 186 7.4.6 PA 6: Delivery through Tunnel...... 187 7.5 Summary ...... 188 CHAPTER EIGHT ...... 189 CONCLUSION ...... 189 8.1 Introduction ...... 189 8.2 Key Findings ...... 189 8.2.1 Six Dimensions collectively represent the Last Mile Delivery Constraints .... 190 8.2.2 Last Mile Delivery Impedance varies across Metropolitan Melbourne ...... 190 8.2.3 Future of Last Mile Delivery is shaped by Infrastructure Supply and Landuse Intensity...... 191 8.2.4 Stakeholders with Power and Interest can cause Positive Change in Last Mile Delivery ...... 191 8.3 Addressing the Research Questions ...... 192 8.4 Contribution of the Study ...... 193 8.4.1 Theoretical Contribution ...... 193 8.4.2 Methodological Contribution ...... 194 8.4.3 Planning Implications ...... 195 8.5 Limitations ...... 196 8.6 Future Research Direction ...... 197 8.7 Final Concluding Remark ...... 198 REFERENCES ...... 200 Appendix A: Ethics Approval ...... 231 Appendix B: Scenario Thinking Workshop Worksheet ...... 241 Appendix C: Location of Activity Centres ...... 243 Appendix D: Planning Zones within Metropolitan Melbourne ...... 245

viii

LIST OF TABLES

Table 1.1: LMD in Australian cities between 2001 and 2016 in billion tonne kilometre ...... 12 Table 1. 2: Population growth and last mile logistics trend in Melbourne ...... 14

Table 2.1: Various descriptions of city logistics ...... 25 Table 2.2: Definitions of last mile delivery ...... 26 Table 2. 3 : Descriptive terms applied to urban street network ...... 39

Table 3.1: Thinking process tools ...... 50 Table 3.2: Summary of basic approaches in Scenario Thinking ...... 54 Table 3.3: Approaches in the study of the future ...... 57 Table 3.4: Built Environment, Planning System and Transport System Constraints ...... 60 Table 3.5: Population density and distance of Councils from CBD ...... 62

Table 4.1: Paradigm and ontology, epistemology and methodology...... 76 Table 4.2: Amalgamated cities to form the current 31 councils of Metropolitan Melbourne . 81 Table 4.3: Local Councils within Metropolitan Melbourne ...... 82 Table 4.4: Key case study area comparison ...... 86 Table 4.5: Constraints and dimensions of data ...... 97

Table 5. 1: Key constraints identified by the workshop stakeholders ...... 117 Table 5.2: Built environment, planning system and transport system constraints...... 118 Table 5. 3: Road segment with level of impedance ...... 120 Table 5.4: Number of road network segments and impedance rank ...... 123 Table 5. 5: Level of impedance comparison across the inner, middle and outer rings ...... 127 Table 5.6: ANOVA table across Metropolitan Melbourne ...... 129 Table 5. 7: Small area analysis ...... 136 Table 5. 8: Metropolitan Melbourne and small area level of impedance ...... 136

Table 7.1: Desirable Distribution Objectives ...... 168

LIST OF FIGURES

Figure 1.1: Population growth of Australia compared with other countries of the world ...... 7 Figure 1.2: Last Mile delivery and population growth in Australia ...... 8 Figure 1.3: Annual last mile delivery space requirements between 2015 and 2020 ...... 8 Figure 1.4: Population projection and last-mile delivery...... 10 Figure 1.5: LMD in Australian cities between 2001 and 2016 in billion tonne kilometre ...... 13 Figure 1.6: A comparison of Melbourne population and LMD...... 14 Figure 1.7: Last mile delivery hotspots within the Melbourne CBD ...... 15 Figure 1.8: Five staged research methodology ...... 19 Figure 1.9: Thesis Structure ...... 20

Figure 2.1: Hierarchical structure of urban freight ...... 24 Figure 2.2: The last mile in inland freight distribution...... 28 Figure 2.3: A typology of Last Mile Delivery system ...... 29 Figure 2.4: Themes and subthemes of SCM ...... 31 Figure 2.5: Transportation, Urban Form and Spatial Structure ...... 36 Figure 2.6: Urban street pattern archetypes ) ...... 40

Figure 3.1: Key questions linking the Theory of Constraints and Scenario Thinking ...... 44 Figure 3.2: Theoretical Framework for linking last mile delivery with urban structure and planning controls...... 46 Figure 3.3: Logistics paradigm ...... 47 Figure 3.4: Embedded methodologies in Scenario Thinking ...... 57 Figure 3.5: Inter-locking interactions between planning system, transport system and last mile delivery within the built environment ...... 59 Figure 3.6: Land area, population and population density of Local Councils in Metropolitan Melbourne...... 63 Figure 3.7: Toll Roads in Metropolitan Melbourne ...... 65 Figure 3.8: Typical Melbourne road - Westgate Freeway ...... 69 Figure 3.9: Shared Tram and vehicle roadway ...... 70 Figure 3.10: Railway boomgate in operation ...... 71 Figure 3. 11: Clayton Road overhead rail pass ...... 71 Figure 3.12: Poath Road, Hughesdale Rail overhead ...... 71

Figure 3. 13: A typical cut out bicycle lane ...... 72

Figure 4.1: Range of methodology and related paradigms ...... 75 Figure 4.2: Melbourne and Suburb (1868) ...... 79 Figure 4.3: Melbourne Population growth 2001 to 2016 ...... 79 Figure 4.4: Metropolitan Melbourne within Victoria ...... 80 Figure 4.5: Metropolitan Melbourne in three rings ...... 82 Figure 4.6: Spatial Configuration of Activity Centres...... 83 Figure 4.7: Location of Activity Centres within Metropolitan Melbourne ...... 84 Figure 4.8: Location of Maribyrnong, Monash and Nillumbik within Metropolitan Melbourne...... 86 Figure 4.9: Seamless integration of Principles of Theory of Constraints and Scenario Thinking ...... 89 Figure 4.10: Interaction and interest of stakeholders to cause the change in last mile delivery ...... 91 Figure 4.11: Scenario Thinking steps and associated questions ...... 93 Figure 4.12: Research Process ...... 95 Figure 4.13: Population Density ...... 98 Figure 4.14: Planning zones ...... 98 Figure 4.15: Traffic Count within Metropolitan Melbourne...... 99 Figure 4.16: Location of Activity Centres ...... 99 Figure 4.17: Number of lanes within Metropolitan Melbourne ...... 100 Figure 4.18: Speed limits ...... 100 Figure 4.19: Location of Traffic lights within Metropolitan Melbourne ...... 101 Figure 4.20: Rail lines including Tram Routes within Metropolitan Melbourne ...... 101 Figure 4.21: Typical buffer tool in ArcGIS ...... 103 Figure 4.22: Typical Clipping tool procedure ...... 103 Figure 4.23: A Typical GIS overlay function ...... 105 Figure 4.24: Impact and uncertainty matrix on horizontal scale ...... 107 Figure 4.25: Vertical movement of dimensions within the Impact and uncertainty matrix .. 107 Figure 4.26: Four quadrants of scenario building ...... 109 Figure 4.27: Stakeholder power and interest matrix ...... 110

Figure 5.1: Stage 1 of the research design ...... 115

Figure 5.2: Last mile delivery impedance index across road segments ...... 121 Figure 5. 3: Calculated level of impedance within Metropolitan Melbourne...... 122 Figure 5.4: Mean last mile delivery impedance within Metropolitan Melbourne ...... 124 Figure 5.5: Population density and impedance level relationship scattered diagram ...... 125 Figure 5.6: Impedance level and distance from Metropolitan Melbourne CBD relationship scattered diagram ...... 125 Figure 5. 7: Inner, middle and outer rings impedance level comparison...... 128 Figure 5.8: LMD impedance Levels across inner, middle and outer rings...... 128 Figure 5.9: Last mile delivery and tram route interface ...... 133 Figure 5.10: Height limit constraint along Cheltenham Road in Dandenong ...... 134 Figure 5.11: Webster Street Dandenong railway boomgate ...... 134 Figure 5.12: Overhead collision resulting from low height bridge ...... 135 Figure 5.13: LMD impedance levels across different councils ...... 137 Figure 5.14: LMD impedance levels within Maribyrnong City...... 138

Figure 6.1: Research Design Stages...... 141 Figure 6.2: Key constraints within the Freight Infrastructure dimension ...... 143 Figure 6.3: Private cars parking within loading zone ...... 144 Figure 6.4: Key constraints within the Landuse Intensity dimension ...... 145 Figure 6.5: Key constraints within the Infrastructure Supply dimension ...... 147 Figure 6.6: Key constraints within the Intersection Controls dimension...... 148 Figure 6.7: Key constraints within the Infrastructure Sharing dimension ...... 149 Figure 6.8: Typical bicycle lane cut out from existing road network ...... 150 Figure 6.9: Key constraints within the Human Behaviour dimension ...... 150 Figure 6.10: Impact -Uncertainty matrix – higher order factor...... 152 Figure 6.11: Scenario scoping – broad descriptions of future scenarios ...... 155 Figure 6.12: Best (A)/Worst (B) scenario ...... 156 Figure 6.13: Best (A)/Best (A) scenario...... 157 Figure 6.14: Worst (A)/Worst (B) scenario ...... 158 Figure 6.15: Worst (A)/Best (B) scenario ...... 159 Figure 6.16: Stakeholder matrix – stakeholders to cause change in last mile delivery ...... 160

Figure 7.1: A Strategic Framework for an efficient, sustainable and cost effective LMD .... 166 Figure 7.2: Strategies, key desirable distribution objectives and applicable actions...... 167

xii

Figure 7.3: Applicable potential outcomes and desirable distribution objectives for policy intervention in LMD ...... 169 Figure 7.4: A localised spatial plan to augment the efficiency of last-mile delivery in Maribyrnong City ...... 170 Figure 7.5: Landuse zoning strategy and associated actions ...... 172 Figure 7.6: Example of curb-side cut-out ...... 172 Figure 7.7: Last mile corridor strategy and applicable actions ...... 174 Figure 7.8: Actions to mitigate LMD challenges through Distribution Network Strategy ... 176 Figure 7.9: Actions to achieve the objectives of the multimodal land transport strategy ...... 177 Figure 7.10: Action to achieve the outcome objectives associated stakeholder Engagement in Decision Making ...... 178 Figure 7.11: Shift of interest of federal, state and local government and Transurban to cause change in last mile delivery ...... 179 Figure 7.12: Mobile UCC (Source: Straightsol project 2015) ...... 182 Figure 7.13: Smile Project typology in Barcelona ...... 183 Figure 7.14: Typical operation of Consolidation ...... 183 Figure 7.15: Information Technology and Crowd sourcing in last mile delivery ...... 185 Figure 7.16: Use of drones for delivery ...... 186 Figure 7.17: Typical use of tunnel for delivery to city centre ...... 187 Figure 7.18: Typical use of tunnel for delivery to city centre ...... 188

xiii

LIST OF ABBREVIATIONS

ABS Australian Bureau of Statistics

AHURI Australian Housing and Urban Research Institute

B2B Business to Business

B2C Business to Customer

BB Best-Best

BESTUFS Best Urban Freight Solutions

BW Best –Worst

C2C Consumer to Consumer

CBD Central Business District

CIA Cross Impact Analysis

CLFQP Central London Freight Quality Partnership

CRT Current Reality Tree

CSCMP Council of Supply Chain Management Professionals

DDO Desirable Distribution Objectives

EC Evaporative Cloud

ESRI Environmental Systems Research Institute

EU European Union

FRT Future Reality Tree

GIS Geographic Information System

GPS Global Positioning System

HCV Heavy commercial Vehicle

HI High Interest

HP High Power

IPCC Intergovernmental Panel on Climate Change

xiv

ITS Intelligent Transport Systems

LCV Light Commercial Vehicle

LI Low Interest

LM Last Mile

LMD Last Mile Delivery

LMDCS Last Mile Delivery Corridor Strategy

LML Last Mile Logistics.

LP Low Power

LUZS Land-use Zoning Strategy

NGO Non-Governmental Organisations

OECD Organisation for Economic Co-operation and Development

PTV Public Transport Victoria

SCM Supply Chain Management

SMILE Smart mobility & last MILE logistics

ST Scenario Thinking

TCE Transaction Cost Economic

TOC Theory of Constraints

TIA Trend-Impact Analysis

TP Transit Point

TSM Transport System Management

UCC Urban Consolidation Centres

UDC Urban Distribution Centre

UNDESAPD United Nations Department of Economics and Social Affairs

VPP Victoria Planning Policy

WB Worst-Best

WW Worst-Worst

LIST OF PUBLICATIONS

Journal papers

1. Ewedairo, K, Chhetri, P. and Jie F. (2018). “Estimating transportation network impedance to last-mile delivery: A Case Study of Maribyrnong City in Melbourne", The International Journal of Logistics Management, Vol. 29 Issue: 1, pp.110-130. ABDC Ranking: A.

Refereed Conference Papers (Full Paper)

1. Ewedairo, K, Chhetri, P. Dodson, J and Rahman S. (2018). “Building Last Mile Delivery Scenarios: A Case Study of Melbourne”, International Conference on Industrial Engineering and Engineering Management (IEEM), 16 - 19 December, Bangkok, Thailand.

2. Ewedairo, K, Chhetri, P. and Jie, F. (2016). “Estimating last-mile delivery impedance: a city logistics challenge”, 21st International Symposium on Logistics (ISL), 3-6 July, Kaohsiung, Taiwan.

3. Ewedairo, K, Chhetri, P. and Dodson, J. (2015). “A GIS methodology for estimating the transport network impedance to last-mile delivery” 7th State of Australian Cities Conference, 9-11 December, Gold Coast, Australia.

xvi

ABSTRACT

‘Last mile’ in cities is not merely a logistics problem, but also a significant urban planning challenge. Last mile is the final leg of the supply chain that involves high-frequency, low- volume, and short-haul distribution of goods to end consumers. This last leg is the most important, yet the least efficient part of supply chain. With rapid growth of online retail transactions and increased supply chain complexity of the globalised production networks, the size, scale and scope of last mile-driven logistics problems are most likely to escalate in the immediate future. Transport infrastructure and planning controls are among the key factors that contribute to the severity of last mile delivery (LMD) problems in large cities. However, interrelationships and interdependencies between last mile logistics and transport networks and urban planning controls are neither theoretically evaluated nor empirically tested in the extant literature.

This thesis aims to measure and map the potential transportation network impedance to last mile delivery, build future last mile delivery scenarios and formulate a strategic framework to mitigate the risk of last mile delivery failure within Metropolitan Melbourne. A five-stage Scenario Thinking method was applied to understand and analyse the provision of last mile delivery and the associated ‘critical uncertainties’. A Geographic Information System (GIS) was used to compute and visualise the potential transportation network impedance to last mile delivery within the Metropolitan Melbourne. Spatial data representing the key constraints of transportation network and planning controls were used to compute the potential transportations network impedance to last mile delivery.

A Scenario Thinking (ST) workshop was conducted with 14 participants who represent three major stakeholders, namely operators, administrators and users. This was designed to discuss and identify key planning and transportation constraints contributing to last mile challenges and to formulate future possible and plausible last mile delivery scenarios. Thirty- four major issues that underpin last mile logistics were identified, which were clustered into six thematic dimensions through an iterative consultation process. These include: i) Freight Infrastructure; ii) Infrastructure Supply; iii) Landuse Intensity; iv) Infrastructure Sharing; v) Intersection Controls and vi) Human Behaviour. Infrastructure Supply and Landuse Intensity were found to represent higher uncertainty and higher impact on city logistics provisions. Hence, they were used to build future LMD scenarios.

xvii

Spatial data of seventeen mappable constraints, including traffic count, population density, zoning, proximity to Melbourne CBD and activity centres, intersection constraint, speed limit, number of lanes, toll, railway boom gate, traffic lights and trams routes, were standardised and aggregated using a composite index technique. The generated LMD impedance index was then mapped using an overlay function to estimate the potential hindrance to last mile delivery as imposed by built and regulatory environment. The mapped outputs illustrate significant spatial variations in LMD impedance levels across different parts of Metropolitan Melbourne. Impedance to last mile reduces with increased distance from the Central Business District, found to be high within the designated activity centres and varies across inner, middle and outer rings.

The results reveal that the future outcomes of last-mile delivery is dependent upon which scenario eventuates out of the four scenarios formulated using Infrastructure Supply and Landuse Intensity. The best-best scenario characterises an increase in infrastructure usage, decrease in road congestion, lower delivery cost, and increased last mile delivery productivity and efficiency; whilst the worst-worst scenario represents a decrease in the usage of logistics infrastructure, transport delay and congestion, increased last-mile delivery cost, loss of work productivity, reduced deliveries per day and higher environmental impacts.

Further analysis indicates the relative positioning of major stakeholders based on an interest- power matrix. Last mile logistics stakeholders such as Federal, State and Local Governments, and Transurban with high-power tend to exhibit low-interest in provisioning last-mile logistics; while those with high interest such as Drivers and Business owners have low power to influence any positive change to enhance the efficiency of last-mile delivery. Stakeholders with high-power and high-interest were identified to include Public Transport Victoria, VicRoads, Trader Association, Port Authority who could assist in policy-making to help improve last-mile delivery efficiency and curtail carbon footprint within an acceptable level.

In this study, a strategic framework is developed to support the formulation of key objectives and future-oriented actions. Five strategies along with key actions were proposed to tackle the LMD challenges associated in Melbourne. These include: land-use zoning, last mile corridor, distribution network strategies, multimodal use strategy and stakeholder engagement strategy. The land use zoning strategy can be implemented to geographically demarcate last-mile delivery zones to improve the efficiency of last-mile delivery to retail businesses within the localised area. The last mile corridor strategy would expedite last-mile delivery along the

xviii

main arterial networks through the development of linear freight routes to improve last-mile connectivity between key business hubs. This geo-targeted strategy will help reduce the environmental footprint of last-mile delivery, ease traffic bottlenecks and potential conflict between last-mile delivery trucks and commuters by confining truck movements to designated routes. The distribution network strategy would promote consolidation of goods through a holistic integration of people, facilities and transportation infrastructure as a single unified city logistics network to support the development of Urban Distribution and Consolidation Centres in vicinity to the Activity Centres. The Multi modal use promotes LMD using trucks and train, trams or bicycles as an integrated system. The stakeholder engagement strategy places a greater emphasis on the shift in interest of stakeholders with power to cause positive change through advocating for the implementation of effective policies and regulations.

Overall, this thesis is the first study that applied the Scenario Thinking method along with GIS to construct, measure and map the potential last-mile delivery impedance. Scenarios were formulated to provide improved understanding of future last-mile delivery. Strategies were recommended to help develop operational plans and deploy future investment to tackle the challenges associated with the worst/worst scenario. The mapped outputs improve the future understanding of the complex interactions between transportation infrastructure, planning controls and last-mile delivery. This understanding and proposed strategies in turn would help enhance the efficiency of last-mile delivery in the context of large cities.

xix

CHAPTER ONE

INTRODUCTION

1.1 Introduction

The ‘last mile’ delivery in cities is not merely a logistics problem, but also a significant city planning challenge. Last mile delivery are expected to escalate because of increased online retail transactions, changes in demand for global products at local and overseas markets resulting from complex and dynamic population and demand changes. Online retail transaction for example has resulted into increased business to customer last mile delivery. In addition, growing demand for a greater variety of goods by consumers, reduction in life cycle of products and limited capacity of warehousing sales floor are exerting pressure of time and space for deliveries in a highly dynamic urban environment. The changes resulting from city compactness (to curb urban sprawl) further intensify the complexity of last mile logistics, with the transportation networks and planning regulations contributing to the urban dynamics and supply chain complexity.

Last mile logistics is simply a part of the supply chain and an integral component of city logistics. It ensures that the process of getting goods to the consumer is completed. The term “last mile” is used for the last leg of the delivery process. It is also significant to the maintenance of urban liveability, quality of life and the functioning of urban economy. In the last decade, the role and functions of last mile delivery (LMD) have become vital given the need to get goods to the final customer within the constraint of highly restricted time-window in the urban setting. In addition, LMD is one of the physical and business activities perceived by end users to link with environment impact and user conflicts in their immediate environment. The conflicts necessitate imposition of different regulatory regimes and network constraints to control the timing, location, scale and intensity of last mile delivery vehicles which impede the delivery process. The LMD impedance creates operational inefficiency and transportation delay of varying degrees across the city and makes last mile delivery costlier. With this growing complexity, the size, scale and scope of last mile-driven logistics problems are most likely to escalate in the immediate future. Interrelationships and interdependencies between last mile logistics and transport networks and urban planning

controls therefore need to be theoretically evaluated and empirically tested to provide deeper insights into the future.

This chapter provides an overview of this study. The chapter is organised into eight sections. Following the introduction, Section 1.2 provides the problem statement, which highlights the challenges of growing complexity of city logistics operations in a dynamic urban environment. Section 1.3 presents the rationale for the research, while section 1.4 describes the scope of the study, which is business to business (B2B) last mile delivery. Section 1.5 presents the research aim and questions, while section 1.6 presents a synopsis of research methodology. Section 1.7 outlines the thesis structure and section 1.8 concludes with the summary of the chapter.

1.2 Problem Statement

The final leg in the supply chain process is often known as last mile, which is deemed as the most inefficient, costly and cumbersome component of freight logistics. This inefficiency in last mile operations is further worsened through different levels of impedance produced by (i) increased population density and compactness, (ii) the built and regulatory environment (iii) lack of commensurate infrastructure development and (iv) changing consumer demand and growing e-commerce. The built environment continues to change and the regulatory environment - through transportation and planning systems controls – becomes more stringent to response to changing planning paradigm. These changes will continue as the city gets more compact. How last mile delivery will be affected by these changes is currently unknown. The challenge to last mile delivery from these changes therefore deserves to be investigated to provide clarity and strategic direction.

The need for business to business last mile delivery are expected to continue to grow due to changes in demand for global products at local and overseas markets resulting from globalised production and supply networks to fulfil dynamic and uncertain demand from goods at lower prices. This is further heightened by reconfiguration of local distribution networks which characterise reduction in warehousing space, greater spatial dispersion of retail outlets and shifting of production to overseas markets. Consumers are requesting a greater range of goods, both physically through brick and mortar and online through

ecommerce, and they want the goods right on time and require a higher frequency of delivery (Esper et al., 2003).

In Australia, last mile delivery is identified to result to a disconnect in the supply chain as it is a by-product of rises in the sizes and number of vehicles on major routes, or as a result of a disparity between transport planning and landuses (Infrastructure Australia, 2011). Freight Australia specifically identifies that transport infrastructure policies do not adequately address last mile delivery issues and those that exist do handle last mile delivery haphazardly (Infrastructure Australia, 2011).

While the consumers are more likely to procure goods ‘just in time’, via roads, there are several physical and transport controls which affect road transport system including zoning, transport restrictions, speed limits, curfew and parking, reduction or removal of loading and unloading. Few of these controls are implemented in pursuance of a compact city design of Melbourne. The impedance imposed on last mile logistics resulting from the redesign and retrofitting inner and middle city suburbs in Melbourne must be overcome to realise just-in- time delivery orders. Compact city model requires densification of residential and intensification of land use (Jenks et al., 2008). Arguably, the consequence of the pursuance for greater compactness of the city on last-mile is a major hindrance that inhibits efficiency of last mile logistics.

Few policies adopted by local authorities to achieve planning objectives have hindered last mile delivery by placing more stringent restrictions on the provision of city logistics (Anderson 2000; Dablanc 2007; Bliemer et al. 2016). Planners are often perplexed in establishing city logistics projects as they may make other planning objectives hard to achieve (Goldman and Gorham 2006; van Rooijen and Quak 2010; Gammelgaard 2015) Specifically, Bliemer et al., (2016) confirms that planners have since the 1970’s struggled to develop strategies to reach a solution for last mile delivery in cities and concluded that city compactness will further increase the hindrances to last mile delivery.

The study of this LMD impedance and how the future will react to the changes as a result of the increased city compactness is lacking in city logistics literature. As the city gets more compact, infrastructure to meet the requirement of urban freight is unlikely to be achieved on-time or at the optimal price for city dwellers and retailers within the activity centres (Han et al., 2005 and Agunbiade et al., 2014). Despite this dilemma, land-use policymakers

consider transport planning, population growth and spatial spread of city separately or in isolation from each other (Tschopp et al. 2005; Angel et al. 2011). By extension, city logistics characteristics are neglected in policy making process by policy makers.

In addition, the future of last mile is as uncertain as the future of the urban environment. The uncertainty associated with technological, regulatory and the built environment will unquestionably affect the future of last mile delivery. The efficiency of last mile delivery in the near future is heavily dependent on the increasing complexity of the rapidly changing urban environment. Yet no major study has unfolded the complex relationships, interaction and interdependencies between the urban environment and last mile logistics.

The increased globalisation of logistics and supply chain networks also contributes to this dynamics and complexity. Arguably, the future of last mile is as uncertain as the future of the urban environment. The uncertainty in the level of urbanisation dynamics will unquestionably affect the future of last mile delivery.

Resulting from city compactness, roads are getting narrower and more congested. Parking spaces are also getting scarce and hard to find for last mile delivery vehicles, resulting in delays and impacting just in time delivery as well as increased cost of delivery. Conservatively, illegal parking of delivery parcel vehicles is the third leading cause of cities congestion behind crashes and construction (Han, et al., 2005).

Furthermore, given the size of vehicles involved in city freight logistics, turning and manoeuvring often causes road rage, accidents. In addition, there are often conflicts between city freight operators and other road users (Conway et al., 2013), which would make getting a consensus harder for the local planning authorities due to conflicting planning objectives. A consolidated approach to solve a multi-industry problem is now required to mitigate the potential risk of supply failure and the concomitant potential business closures.

Prior to the beginning of the 21st century, the consideration of freight distribution in urban setting remained limited within the urban planning discussion (Han et al. 2005; Morris 2009; Morris et al. 1999; Browne et al. 2007; Rodrigue et al. 2013). Japan and cities in Western Europe pioneered early application of city logistics (Rodrigue et al., 2013). This is because these cities lack available land necessitated by the form of the cities.

It is clear from above that as population increases and the city continues to get more compact,

the need for last mile delivery will continue to increase. Urban Planners therefore, need to give consideration as to how last mile delivery challenges will be addressed. Thus, new transportation infrastructure or better utilisation of existing infrastructure is required to accommodate the future demand for growing freight tasks in larger cities such as Melbourne. Research on the impact of compactness and future of last mile delivery is currently lagging in the extant literature. This study aims to address this research gap by developing a better understanding of the interplay of built, transport and regulatory regime to help develop strategies to reduce supply risk.

An insight into the future of last mile delivery within dynamics and uncertain urban environment through Scenario Thinking can facilitate strategic thinking. One of the ways to gain a better understanding of the future possible and plausible of last mile delivery is creating situations through Scenario Thinking (Ewedairo, et al., 2018). Rather than forecasting, building last mile delivery scenarios will assist in planning for future last mile delivery in a changing and complex compact city being proposed in Melbourne. Different alternative futures can be contemplated and strategic options (Ewedairo, et al., 2018) can be evaluated against plausible futures (MDS Transmodal Limited, 2012).

Urban freight distribution within the planning discipline must now be integrated to develop city-wide planning to deal with challenges operating within an interrelated and interdependent city system (Rodrigue et al., 2013). An improved understanding of plausible last mile delivery future scenarios would help transport planners and logisticians to plan future investment strategies and operational plans to deal with the future challenges.

1.3 Research Rationale

This section discusses the rationale for the research based on problem identification in section 1.2. These problems include the expected changes in the built environment resulting from city compactness in response to growing population growth and increases last mile delivery. The need to understand the dynamics of last mile logistics in large cities is growing in recent decades. The effects of the changing patterns of population growth, city compactness, use of Transport System Management (TSM) to manage road infrastructure, as well as planning and regulatory controls on last mile delivery are largely unknown. The rationale behind the undertaking of this study is presented under the following themes of challenges.

1.3.1 Increased Population Growth and Growing Pressure on Cities

The level of urbanisation requiring last mile delivery has increased to more than 54.5% across the world and, notably, it has reached 73% in Europe (United Nations, 2016). The population of Delhi, Sao Paulo, London and Lagos is projected to increase from 38 million, 21 million, 10 million and 13 million in 2016 to 36 million, 23 million, 11 million and 24 million by 2030 respectively (United Nations, 2016). The population increases, and the densification of commercial and residential areas will continue to drive the need for more last mile delivery. Even in countries like Japan, the Baltic Region, Russia and Europe where population has been projected to shrink over the years, what is expected is continuation of increases last mile request (Harrington, 2015).

In the same vein, Australia is experiencing an accelerated population growth with Sydney and Melbourne accounting for greater portion of the total population. Figure 1.1 reflects the growth of Australia between 2001 and 2012 compared with other countries of the world. It reveals that the population of Australia increased by approximately 1.25% in 2001 to 2.1% in 2009 and 1.75% in 2012. No other countries in the report have this level of percentage population increase within the OECD (Figure 1.1).

Figure 1.1: Population growth of Australia compared with other countries of the world (Source: World Bank 2015).

To add to last mile delivery issues is the challenges associated with burgeoning urbanisation that transcend over merely providing housing and employment to growing population but to the fulfilment of growing consumer just-in-time demand which necessitates supplying materials and consumables to most inaccessible, restricted and time-sensitive urban spaces.

Figure 1.2 shows the relationship between population growth and last mile delivery in Australia between 2001 and 2014. While the growth of population is insignificant, yet the last mile delivery shows a significant increase in relative term. In addition, as shown in Figure 1.3, Australia has ranked fifth in the annual logistic space, followed by China, United States, India and UK/Ireland.

250

200

150 Last Mile Delivery Population 100

0 20012002200320042005200620072008200920102011201220132014

Figure 1.2: Last Mile delivery and population growth in Australia Source: BITRE (2015) and ABS (2015).

Figure 1.3: Annual last mile delivery space requirements between 2015 and 2020 (Source: Harrington, 2015)

1.3.2 The Changing Built and Regulatory Environment under the Compact Cities Model

As a result of the growth in population, Victorian government seeks to curb outward urban sprawl to concentrate on establishment of compact cities. Chhetri et al., (2013) highlight the general purpose of city compactness model to include reduction in urban sprawl, decrease in car ownership (Davison, 2004) and greater utilisation of existing infrastructure and services in more established areas of inner and middle-ring suburbs and around high intensity Activity Centres (Department of Infrastructure, 2002). Due to the increased compactness of inner suburbia, higher population densities in central areas have grown, roads are getting narrower, loading and unloading spaces are getting reduced or totally removed, existing transportation infrastructure is shared by bus lanes and bicycle paths through the Transport System Management (TSM) initiatives. These changes have resulted in increased road congestion, and consequently delays in last mile delivery and loss of productive time for the workforce and operators.

Scholars have over time identified loading and unloading as one major necessity for city logistics efficiency and effectiveness (Ogden 1992; Visser et al. 1999; Taniguchi 2001; Taniguchi et al. 2014; Pivo et al. 2002; Figliozzi 2009; Taniguchi and Tamagada 2005; Dasburg and Shoemaker 2008; Allen et al. 2012; BESTUFS II 2007; Comi and Conte 2011). The type and availability of such loading and unloading operations can impact the overall process of city logistics. Loading and unloading is a major problem within the inner-city suburbs but might not be necessarily a problem in middle or outer suburbs.

In addition, it has been established that city planners and policymakers considered population growth and spatial spread of city separately or in isolation from each other (Agunbiade et al., 2014). By extension, last mile logistics operations are neglected in policy making process by policy makers (Tschopp et. al. 2004; Angel et. al. 2011) in the discussion of the consequences of implementing city compactness model.

There is also growing demand anticipated for last mile delivery by 2040 (Harrington, 2015). Figure 1.4 indicates the impact of population increase on last mile delivery requirements across the world (Harrington, 2015). While many of the countries will experience moderate to higher population growth, shrunk population is projected for Japan, the Baltic Region, Russia and South East Europe among others. Despite this shrink, last mile requirements will continue to increase in these countries. Hence the last mile delivery challenges will increase

and a bigger and more robust logistics and distribution networks across the cities are required to support vast increases in online demand for urban freight. For example, Figure 1.3 reveals that Tokyo in Japan is projected to experience a population shrink, though it will require greater number of last mile delivery per household in comparison to other cities (Figure 1.4). Australia will require approximately 3.17 million square metres of annual last mile space between 2015 and 2020.

Figure 1.4: Population projection and last-mile delivery (Source: Harrington, 2015)

1.3.3 A Mismatch between Consumer Demand for Goods and Transportation Infrastructure Supply

Compactness, increased car ownerships and inadequate transportation infrastructure result in an increased road congestion, which translates into freight delivery delays and loss of productive time for last mile delivery vehicles. In some cases, loading areas or manoeuvring space are removed to create bicycle paths resulting in to illegal parking of last mile delivery vehicles due to the lack of manoeuvring space.

Han, et al., (2005), estimated that in large urban areas in the U.S, total delay caused by illegally parked pickup/delivery vehicles was 500 million vehicle hours annually costing $10 billion in lost time. They concluded that there has been little systematic planning effort to reduce truck-related congestion in urban areas or addressing loading and unloading of last

mile delivery vehicles. Especially when it comes to studies of built environment factors such as land use, streetscape, and urban design, last mile delivery is not considered a factor in the planning process. Hence, repeated cycling of the last mile trucks can be as a result of lack of parking spaces (curb space) or insufficient off-loading facilities (Morris, 2009), or lack of manoeuvring space as a result of the compact city (Bliemer et al., 2016).

With population increase, there has been a continued growth in e-commerce with ease in making purchases through the internet and shift from brick and mortar purchases. This has generated enormous requests for last mile delivery of goods and services from the suppliers’ side. The number of consumer related trips are therefore being substituted by suppliers’ related last mile delivery. However, necessary transportation infrastructure fails to match with the rapid population growth within the changing urban environment. The use of Transport System Management (TSM) technique to increase the capacity of road infrastructure without increasing the size of the roadways to tackle the negative impact of urban densification on last mile delivery is found to be less effective. Furthermore, the trend in the reduction in warehouse capacity has also contributed to the frequency in request of goods from fresh food to large furniture. All these changes have resulted in further increase in last mile demand frequency in business to business (B2B) and business to consumer (B2C) deliveries as consumers now requires quick and just on time delivery, in some cases 24 hour delivery (Visser et al., 2014)

1.3.4 Changing Consumer Demand and Distribution Methods

With the growth and changing patterns of consumers’ needs, preferences and choices, there are requests for a greater variety of goods by consumers, with reduction in life cycle of products (Geyer et al, 2007) and restricted size of warehousing trade floor (Greasley and Assi, 2012). The changes resulting from city compactness in recent decades (to curb urban sprawl) increased complexity of last mile logistics, while supply chain networks are contributing to this urban dynamics and supply complexity.

City logistics will continue to grow due to increased changes in demand by increased population concentrations in strategic nodes of economic vibrancy. Also, consumers continue to request for a greater variety of goods, ranging from perishable goods (food) to larger household items, both physically via retail outlets and online. Consumers want the goods right on time and in most cases in a higher delivery frequency. The request for last

mile delivery will therefore continue to grow in the future as the consumers’ preferences, predilections and choices change and Metropolitan Melbourne continues to grow in shape and size (Stratec, 2005).

In addition, last mile delivery has continued to increase in terms of volume, distances and subsequent fuel consumption (Ewedairo, et al., 2016). These developments, together with the need for an agile, lean and just-in-time logistics, have resulted in a growth of last mile delivery problems (Stratec, 2005).

There are corresponding increases in the number and volume of last mile delivery to increases in population in Australian cities. Table 1.1 and Figure 1.5 reveal the changes in last mile delivery in terms of billion tonne kilometre in Australian cities between 2001 and 2016 (BITRE, 2017). These figures show an increase in last mile with the exception of 2008 during the financial crises. Specifically, Table 1.2 and Figure 1.6 illustrate the increase in volume of last mile delivery in Melbourne and population growth and there exist high correlation between population growth and last mile delivery. The data show that percentage growth of last mile is higher than population growth in Melbourne.

Table 1.1: LMD in Australian cities between 2001 and 2016 in billion tonne kilometre

Year Sydney Melbourne Brisbane Adelaide Perth Hobart Darwin Canberra 2001 9.9 9.9 5.8 2.1 3.7 0.3 0.2 0.4 2002 10.2 10.2 6.2 2.2 3.8 0.3 0.2 0.4 2003 10.4 10.4 6.6 2.3 4.1 0.3 0.2 0.5 2004 10.6 10.6 7.0 2.4 4.2 0.3 0.2 0.5 2005 10.9 10.9 7.2 2.5 4.6 0.4 0.3 0.5 2006 11.0 11.1 7.7 2.7 5.1 0.4 0.3 0.5 2007 11.2 11.4 8.1 2.7 5.4 0.4 0.3 0.5 2008 10.9 11.2 8.0 2.7 5.4 0.4 0.3 0.5 2009 11.0 11.3 8.2 2.7 5.5 0.4 0.3 0.5 2010 11.3 11.6 8.6 2.8 5.7 0.4 0.3 0.6 2011 11.7 11.9 9.0 2.9 6.0 0.4 0.3 0.6 2012 12.0 12.1 9.1 2.9 6.4 0.4 0.3 0.6 2013 12.3 12.4 9.2 3.0 6.6 0.4 0.3 0.6 2014 12.6 12.7 9.4 3.1 6.9 0.4 0.3 0.6 2015 13.0 13.0 9.7 3.1 7.0 0.5 0.3 0.6 2016 13.4 13.4 10.0 3.2 7.3 0.5 0.3 0.6 Source: BITRE (2017)

Sydney 10 Melbourne Brisbane 8 Adelaide 6 Perth Hobart 4 Darwin Canberra 2

0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Figure 1.5: LMD in Australian cities between 2001 and 2016 in billion tonne kilometre

Specifically, between 2011 and 2016, the population of Melbourne rose from 4.19 million to 4.85 million, an increase of 10.47%; while the last mile delivery increased from 11.9 billion tones to 13.4 billion tones and increase of 12.6% (BITRE, 2017; ABS, 2018) (Table 1.2 and Figure 1.5). From this, Melbourne is experiencing accelerated population growth as well as the growth in urban freight within the city. This change could be attributed to increased migrant population, changing growth area boundary to accommodate population demand in affordable suburbs, creation of new activity centres to increased housing density and varied landuse, as well as changing planning and transportation regulations.

Table 1. 2: Population growth and last mile logistics trend in Melbourne

Last Mile Delivery Year Population (million) (billion tone) 2001 3.33 9.9 2002 3.35 10.2 2003 3.55 10.4 2004 3.59 10.6 2005 3.63 10.9 2006 3.37 11.1 2007 3.80 11.4 2008 3.89 11.2 2009 3.99 11.3 2010 4.07 11.6 2011 4.19 11.9 2012 4.24 12.1 2013 4.34 12.4 2014 4.44 12.7 2015 4.45 13.0 2016 4.85 13.4 Source: BITRE (2017) and ABS (2018)

16 6

14 5 12

4 10

8 3

6 2 4 1 2

0 0

Population (million) LMD (Billion per tone)

Figure 1.6: A comparison of Melbourne population and LMD.

The Melbourne City Council breaks the delivery further to vehicular traffic and delivery for 2012. The breakdown reveals that 3,872 vehicles make last mile delivery per day within the Melbourne CBD. These comprise 3,020 cars/vans and 851 trucks, representing 77.1% of

car/van to 21.9% of diesel driven trucks. Out of these deliveries, 35.7% are likely to be accessed from Victoria Street with William Street deliveries accounting for the highest single last mile vehicular deliveries (Figure 1.7).

Figure 1.7: Last mile delivery hotspots within the Melbourne CBD

The costs of last mile city logistics are estimated at 20 to 30% of all vehicles kilometres (Christopher, 2016). Goodman (2005) has estimated it to account for 28% of all transportation within the supply chain process. Survey by the Australian Bureau of Statistics (2018) indicated that between 2001 and 2014, the volume of last mile delivery by road increased from 139 billion to 286 billion tonne kilometre. This volume is projected to increase by 1.8 times its 2010 level (192 billion tonne) by 2030. By the same period, the Australian population is projected to reach 30 million resulting in increased demand for goods and increased last mile delivery.

Taking this further, light delivery/service vehicles and heavy commercial high impact vehicles entering the Melbourne CBD are estimated to account for 32% and 34% of total city logistics last mile delivery vehicles respectively (Casey, et al., 2014). Despite the higher percentage of high impact vehicles associated with urban freight, local and state governments in Victoria are implementing the strategies driven by compact cities model which will

increase population within the inner city and around activity centres (Department of Infrastructure, 2008) in order to contain and manage urban growth in the outer suburbia of Melbourne (Chhetri et al., 2013).

The extant literature in the field of urban planning or city logistics provides no clear understanding of how these changes outlined above will shape the future of last mile logistics. The questions emerging from above are: does city compactness significantly increase last mile delivery impedance? and how will compactness affect last mile delivery in the future?

1.4 Scope of Study

The scope of this thesis is on the delivery part of last mile logistics with a focus on business- to-business delivery of goods within cities. It excludes the behavioural aspects or operational aspects of last mile logistics. Here the emphasis is placed on the understanding of the constraints on city logistics produced by the built and regulatory environment. In addition, business to customer last mile logistics is outside the ambit of this research. For instance, customer willingness or ability to pick up a delivery from specialised boxes placed at different locations at a convenient time or use a crowdsourcing method (e.g. BringBudy) is not investigated in this study.

Trucks are the only method currently used to physically distribute B2B urban freight in the last mile within Metropolitan Melbourne. The built and regulatory environment, particularly the inner-city areas and key precincts, is more likely to affect the business to business delivery in comparison to B2C delivery in the suburbia. Less dense housing, the predominance of single dwellings and low-volume traffic on suburbia streets make the provision of city logistics much easier. Accordingly, B2B forms the main scope of this research. In this study, the B2B last mile logistics includes the distribution of urban freight associated with retailers and distributors, suppliers of groceries, repair parts, and larger items (furniture and electronics etc.).

The scope is also restricted to the measurement of last mile impedance which is confined to the level of road network resistance imposed on B2B deliveries from the point of pick-up to the point of delivery. Other modes of transport such as rail or motor bikes are not part of this study. Impedance is computed using the spatial and measurable constraints of the built and

regulatory constraints. Other non-mappable constraints are excluded from the measurement framework because of the non-quantifiable geometrical properties. Other social and behavioural characteristics of the urban environment are therefore removed from the measurement of last mile impedance.

1.5 Research Aim and Questions

i. What are the key built and regulatory constraints that impede last mile delivery in an urban setting? ii. What are the core dimensions that drive the future of last mile delivery scenarios? iii. What are the likely future outcomes of last mile delivery scenarios? iv. What strategies can be formulated to mitigate the risk of last mile delivery failure?

1.6 Research Methodology

This study is founded on the Theory of Constraints to analyse and map the potential transportation network impedance to last mile delivery within Metropolitan Melbourne and build future last mile delivery scenarios. The thesis proposes localised strategies including a set of actions to help mitigate future last mile delivery problems.

The thesis is divided into eight chapters, which write up the theoretical framework and the implementation of a five-stage Scenario Thinking method to understand the apparent perverse efficient last mile logistics problems.

1.6.1 Study Area

Metropolitan Melbourne will provide a local study area in which the last mile delivery impedance estimation and mapping will be carried out. Spatial data are readily available for the area. The geography of Metropolitan Melbourne also allows the opportunity to divide the area into three different rings for in-depth study. Metropolitan Melbourne represents the

CBD of Melbourne City which is the capital of Victoria and includes local government areas that extend from the Port Phillip Bay coastline to the west, southwards Mornington Peninsular and the Yarra Valley.

The three local areas of inner, middle and outer allow comparison of the level of impedance across the area. Maribyrnong, Monash and Nillumbik Councils are selected from each of the rings. These Councils represents the inner, middle and outer rings respectively.

1.6.2 Methods

In this study, a five-stage Scenario Thinking method is adopted to build the future last mile delivery scenarios. The five-stage Scenario Thinking method adopted in the thesis is illustrated in Figure 1.8 below. The stages answered the question ‘what to change’; ‘what to change to’ and ‘how to cause change’ Scenario Thinking workshop is conducted to collect data. The thesis used intuitive reasoning in Scenario Thinking, a qualitative approach, to identify key issues and constraints that impede last mile delivery provision.

In stage 1, the key constraints of the built and regulatory environment are identified in the Scenario Thinking workshop with the key stakeholders. At stage 2, these constraints are clustered to generate higher-order dimensions on the basis of commonality. The importance of these dimensions is evaluated in stage 3 for their impact and uncertainty. These are then used to produce the storylines for the development of four future last mile delivery scenarios in stage 4. Stage 5 identifies the major stakeholders in the business of last mile delivery based on power and interest in last mile. The projected outcomes generated for each of the scenarios are the fundamental basis for the generation of key objectives and business strategies with a set of actions to mitigate future last mile delivery.

Primary data are sourced from Scenario Thinking workshop, while spatial data are extracted from VicData and ABS to generate mappable planning and transport network constraints. Mapping and visualisation are carried out with secondary data to depict the spatial variability in last mile delivery impedance.

Identification of last mile delivery impedance constraints Stage 1 What to change? Clustering of Clustering of constraints constraints and defining Stage 2 Defining the cluster outcomes of Dimensions What to change to? Generating Impact/Uncertainty matrix Impact and Stage 3 Uncertainty Index

Scoping the scenarios Scoping and developing the How to cause the Stage 4 LMD scenarios change? Developing the scenarios

Stakeholder Analysis Stage 5

Figure 1.8: Five staged research methodology

Spatial datasets are geo-processed through the different tolls of Geographical Information System (GIS) to generate derived datasets for the research. GIS is a computerised data management system. It is used to capture, store, manage, retrieve, analyse, and display spatial information (Chrisman, 1999). GIS is basically used for so-called ‘supporting tasks’ in transport modelling, with the transportation system of an area being linked to its geography (Maguire et al., 2005). GIS allows the combination of geographical information with other types of information thus, allowing representation of various characteristics of the built environment.

The innovative approach applied in this thesis is the integration of GIS into Scenario Thinking to enable measurement and visualisation of LMD impedance. GIS is applied to compute last mile impedance using spatial constraints of the built and regulatory environment. Scenario Thinking is used to identify the current issues underpinning the efficiency of last mile delivery and to formulate future last mile delivery scenarios. Spatial data are obtained from various sources and are used to extract spatial constraints to represent myriad characteristics of the urban environment. Melbourne is used as a case study area.

1.7 Thesis Structure

This thesis is structured into four parts containing eight chapters. Part 1 contains chapter 1 which sets the context for the research. Part 2, contains chapters 2 and 3. It includes the

background literature and theoretical framework for the research. Part 3 includes chapters 4, 5 and 6. The chapter represents the research interpretation and analysis. Part 4 includes chapters 7 and 8. The chapters focuses on last mile delivery mitigating strategy and conclusion. These parts and associated sections are as illustrated in Figure 1.9.

1 Setting the Context Chapter 1 – Introduction: Organisation of the Thesis, Problem Statement Aim and Research Questions

3 Research Interpretation 2 and Analysis Literature Review and Chapter 4: Research Methodology Theoretical Framework Chapter 5: Identification and Chapter 2: Understanding Last Mile Estimation of Last Mile Delivery Impedance Impedance Chapter 3: Conceptual Frameworks Chapter 6: Building Last Mile

for Building Last Mile Delivery Delivery Scenarios

4 Mitigating Strategy and Conclusion Chapter 7: Strategic Framework for Last Mile Delivery Chapter 8: Conclusion

Figure 1.9: Thesis Structure

Chapter 1 introduces the research topic and provides the justification for this study. It presents an overview of the research, statement of research problem and its scope. Research aim and research questions are set out in this chapter. The chapter also provides an overview of the research methodology.

Chapter 2 focuses on the discussion on city logistics, its challenges and stakeholders. The chapter proceeds to discuss last mile delivery and the urban environment including urban form and last mile delivery. The characteristics and elements of a compact city are discussed in the chapter. The chapter further establishes the interaction between the built and regulatory environments with last mile delivery. The constraints of the built and regulatory environment on last mile delivery are identified in this chapter.

Chapter 3 proceeds to discuss Scenario Thinking method used in the thesis. It explains how Scenario Thinking is applicable within theory of constraint and the thinking process. The chapter establishes the role of theory in any scientific discussion.

Chapter 4 presents the primary research methodology underpinning this study, which includes details of the research paradigm, empirical research design, research analysis, research instrument, data collection process including the procedure of the Scenario Thinking workshop sources of data. Ethical considerations and conclusion leading to the next chapter are also discussed.

Chapter 5 discusses the first stage of the Scenario Thinking and proceeds with the estimation and mapping alongside visualisation of last mile delivery impedance across Metropolitan Melbourne with deep examination at impedance at localised level of Maribyrnong, while Chapter 6 analyses the results of the Scenario Thinking workshop. The chapter provides discussion of the remaining four stages of Scenario Thinking establishing the different dimensions and representing the power and interest of stakeholders in last mile delivery.

Chapter 7 presents the strategies to mitigate last mile delivery problems highlighted in the findings. The chapter discusses the different actions that can be combined to mitigate last mile delivery and proposes specific strategies that can be applied within Metropolitan Melbourne. Chapter 8 summarises the research findings, discusses how the research question are answered and highlights the major contributions and limitations of the research. Finally, suggestions for future research are proposed in the section followed by concluding remarks.

1.8 Summary

This chapter introduced the research topic, identified the problems and provided the justification for undertaking this study on last mile logistics. The chapter presented the scope of study which is specifically B2B last mile delivery. The overarching aim and research questions are also set out in the chapter. The chapter established needs to provide insights into the future of LMD. The chapter also highlights the changing urban environment resulting from the pursuance of compact city to curb urban sprawl and the consequence on last mile delivery. Integration of transport network and planning controls as key drivers of LMD impedance is lagging in current discussion. The urgent need to unfold the future of last mile delivery in a changing urban environment is argued.

The next chapter forms part 2 of the thesis. This part includes Chapters 2 and 3. It reviews the extant literature and develops a theoretical framework for the research.

CHAPTER TWO

UNDERSTANDING LAST MILE DELIVERY IMPEDANCE

2.1 Introduction

This chapter describes the key concepts and perspectives to reflect and theorise last mile delivery impedance within the context of city logistics. City logistics, urban distribution and last mile delivery are commonly used concepts to represent the movement of urban freight. These concepts are interchangeably used in the supply chain and logistics management. It discusses last mile delivery as a part of the city system, which is shaped by built structures and the regulatory environment. While the built structure relates to the arrangement of land use and infrastructure provisions, the location and control of different landuse is under the different regulations.

This chapter is organised into six sections. Section 2.2, following the introduction, presents a discussion on an integrated last mile logistics including a review of literature on city logistics. The section also describes different approaches used to examine and model the functioning of last mile logistics. Section 2.3 examines the impact of urban form on last mile logistics, while section 2.4 discusses network and accessibility in the context of the city logistics. Four different theoretical perspectives are described to examine the last mile logistics in section 2.5. Section 2.6 presents the summary of the key findings drawn from this chapter.

2.2 City Logistics

City logistics is a provision that link the endusers and the logistics systems including freight distribution, and the urban logistics administrations process and policies (Lu and Borbon- Galvez 2012; Rodrique and Dablanc 2017). The commonly used terms to describe city logistics include Urban Goods Movement (McDermott, 1980), Urban Freight Logistics, (Gonzalez-Feliu et al. 2012); Urban Freight (Tipagornwong and Figliozzi 2014; Allen et al. 2011; Stathopoulos et al. 2012; Cherrett et al. 2012); and Urban Logistics (Alho and De Abreu E Silva, 2015).

Cardenas et al., (2017) conceptualise the movement of urban freight in a hierarchical structure, from macro level (city logistics) through meso level (urban distribution) to micro level (last mile logistics). At the macro level, city logistics relate to vehicle and goods flow and goods characteristics, while at the meso level, urban good distribution focuses on how goods are distributed within urban areas (Fernandez-Barcelo and Campos-Cacheda 2012; Cardenas et al. 2017).

The meso level also includes freight transport inflow to the city, the services and amenities used for freight de-bundling and consolidation, as well as the monetary implications of the activities, and the procedures undertaken by the administrators to manage the urban freight transport. These also include the impact of freight on traffic flows and urban liveability (Crainic et al. 2004 and Browne et al. 2012). Last-mile logistics is at the micro level and is the last leg of urban freight, linking distribution networks with end consumers (Cardenas et al., 2017). However, these terms have often been inconsistently used (Wolpert and Reuter, 2012) to describe the movement of goods in cities (Figure 2.1). Extant literature on city logistics highlights the interchangeable use of these words to represent city logistics (Alho et al. 2014; Cardenas et al., 2017).

Macro level Logistics as a system in the urban City Logistics context - policies, actors, norms, resources, performance, etc.

Meso level Urban Goods Network design, logistics services Distribution and infrastructure alternatives

Micro level Logistics operations and Last optimisation Mile

Figure 2.1: Hierarchical structure of urban freight (Source: Cardenas et al., 2017)

Hicks (1977, pp 101) presented the first formal definition of urban freight, which is currently termed ‘city logistics’. It is defined as “all journeys into, out of, and within a designated urban area by road vehicles specifically engaged in pick-up or delivery of goods (whether the vehicle is empty or not), with the exception of shopping trips” (Lindholm, 2013: pp. 6). Since then, numerous definitions have been presented, with varying levels of detail as presented in Table 2.1.

Table 2.1: Various descriptions of city logistics

Authors Definitions Hicks (1977) “All journeys into, out of, and within a designated urban area by road vehicles specifically engaged in pick-up or delivery of goods (whether the vehicle be empty or not), with the exception of shopping trips” (pp.101). McDermott (1980), “Comprising all freight distributions within an urban area of a city with a population of more than 50,000” (pp. 34). Taniguchi and Thompson “The process for totally optimising the logistics and (2002) transport activities by private companies in urban areas while considering the traffic environment, its congestion, safety and energy savings within the framework of a market economy” (pp. 45). Barceló, et al., (2005) “Freight transport in urban areas and specifically the freight flows associated to the supply of goods to city centres” (pp. 326). Bektas et. al., (2016) “Encompasses the routing and movement of freight across all transport modes, as well as associated activities such as warehousing and exchanging information for the management of freight at each end of its journey” (pp. 154). Dablanc (2007) “Any service provision contributing to an optimised management of the movement of goods in cities” (pp. 284). Crainic, et al., (2009) “Concept that tries to optimise urban freight transport systems by considering all stakeholders and movements in urban areas” (pp. 434). Ehmke (2012) “The second leg of the logistics network including pickup and delivery of goods to customers (pp. 11)

2.2.1 Last Mile Delivery

Last mile logistics involves delivery activities procedure from the last transit point of goods to the final drop point within the supply chain process (Lindner, 2011) as fast as possible. As the last portion of the supply chain (Gevaers et al., 2014), last mile logistics involves de-bundling

of large container load into smaller parcels prior to the final delivery. However, it is necessary to clearly distinguish the delivery part from the entire logistics process (Ewedairo et al., 2018). While last mile logistics includes overall, warehousing, storage and inventory, last mile delivery explicitly relates to the delivery part of the last mile logistics process. In this thesis, last mile delivery excludes logistics processes and operations such as inventory management and scheduling, consolidation and sortation and storage functions.

Last mile delivery is characterised by high-frequency, low-volume short-haul distribution of goods to end consumers over short distances. Anderson et al. (1996), consider last mile delivery as an integral part of last mile logistics and characterise it as the delivery of final products in low volumes and at high frequencies. The different definitions of last mile delivery are captured Table 2.2.

Table 2.2: Definitions of last mile delivery

Authors Definitions Gevaers et al., (2014) The last part of the supply chain (pp. 399). Morganti (2011) “The physical distribution occurring in the last part of supply chain” (pp. 401). Morganti, and Gonzalez-Feliu, Small scale distribution of goods in urban (2015) environment which includes delivery to retailer” (pp. 123). Morris (2009) “Pick-up /drop-off point to the end customer in commercial buildings” (pp. 12). Aized and Srai, (2014) “Final step in business to customer logistics” pp. 1053). One important step in supply chain and business-to- customer paradigms and is responsible for efficient and economical final delivery of goods to customers with final mode of delivery of orders via road (pp. 1060). Ehmke, (2012) “Transportation over short distances with smaller trucks” (pp. 11).

Morris (2009) considered last mile delivery (LMD) to pick up and drop off to end customers in commercial buildings, a business to customer delivery (Aized and Srai, 2014) and economical final delivery of goods to customers. For the purpose of this thesis, LMD is defined as logistics activity (Aized and Srai 2014 and Lindner, 2011), which involves distribution of goods from pickup to drop off point to retailers via road (Morganti 2011b; Morganti and Gonzalez-Feliu 2015; Aized and Srai 2014).

Ewedairo et al., (2018) identified various perspectives of the intricacies of last mile delivery. This includes discussion relating to online retailer deliveries (Esper et al., 2003); online groceries deliveries (Boyer and Hult, 2005); and urban freight consolidation (Browne et al., 2005). Knott (1987) and Morganti (2011a) considered last mile in terms of food delivery with trucks from a distribution centre to a number of camps, while Balcik et al. (2008) examined last mile delivery in humanitarian relief operations from local distribution centres to affected areas.

2.2.2 Characteristics of Last Mile Logistics

The term ‘last mile’ (LM) is commonly used in the field of telecommunications and electronics to describe the final leg of connecting the telecommunication process to the end users (Bhagwat et al., 1996 and Wongthavarawat and Ganz, 2003). In logistics management, last mile refers to the last leg of getting goods to the consumer within supply chain process. Last mile logistics includes all services to get the goods to the final consumer storage and warehousing, sorting, routing and delivery to the final consumer (Mason and Lalwani, 2006). It is used interchangeably with last mile delivery (Goodman, 2005).

Studies in the field of last-mile logistics have investigated the processes and activities within a city (Morris et al. 2003; Ehmke 2012; Gevaers et al., 2014), the mode of delivery (Aized and Srai, 2014) and distances covered (Ehmke, 2012), type of delivery as either B2C and B2B deliveries (Lindner 2011; Harrington et al. 2016). Various issues associated with pickup and deliveries over shorter distances are also highlighted in the existing literature (Morris et al. 2003; Ehmke 2012). Specifically, Morris et al. (2003), examined last mile to deliveries in commercial premises, where the focus remained on B2B deliveries, using smaller trucks (Ehmke, 2012). The frequency of delivery and the distance travel is a function of the span, transportation and physical constraints and population density of cities which are vital to the efficient functioning of last-mile logistics.

Figure 2.2 represents the scope of last mile delivery in the context of the global, and regional supply chain. Its scope is defined on the basis of frequency and capacity of last mile delivery. Generally, the frequency of last mile delivery is high with low capacity of tonnage per delivery compared to low frequency and higher tonnage per delivery at the global level. The frequency of delivery has increased with the greater level of required freight through e- commerce (Smith et al., 2001).

Figure 2.2: The last mile in inland freight distribution. (Source: Rodrigue et al., 2013)

The mass of freight also decreases from drayage (containerised goods) for ocean shipment to cartage (transportation over as short distances) within the city with high frequency. The mode of transportation also decreases from requirement for bigger load unit in a single trip to small load unit that can be effectively transported, otherwise noted as Massification and Atomization respectively (Rodrigue et al., 2013).

Last mile delivery facilitates localised freight distribution to end consumers. Last mile delivery does not take place only in cities, but also in smaller towns or rural areas where goods need to be delivered to the enduser. The challenges associated with those locales are significantly different. The severity of last mile delivery however is much higher in larger cities than regional towns. These deliveries of goods in the last leg of the supply network could be business to business (B2B) (Chopra and Meindl 2001; Morganti et al. 2014; Morganti and Dablanc 2014; Ewedairo et al. 2018) or business to customer (B2C) (Punakivi and Yrjola 2001; Edwards and McKinnon 2010; Janjevic and Ndiaye, 2014; Gevears et al. 2014).

Last mile logistics facilitates movement of goods for transactions associated with Business to Business (B2B), Business to Consumer (B2C) and Consumer to Consumer (C2C) (Temporal, 2005). Figure 2.3 depicts last mile delivery systems types. Type 1 represents pickup and

delivery of goods from warehouses to businesses, type 2, business to business, type 3, business to customer and type 4, customer to customer.

B2B delivery relates to delivery from warehouses to other warehouses, from warehouses to other business. This can be for example from vehicle parts retailer to mechanic workshops or from meat depot to restaurants. Delivery often involves large vehicles like trucks and vans travelling longer distances. On the other hand, B2C are delivery from business to consumers. The delivery involves smaller portion of goods which can be delivered or pickup by the consumer. The C2C typology are transactions between consumers. Overall, the characteristics of each of the last mile typology can be differentiated in the volume, frequency, and vehicle size (Park and & Regan, 2004).

Distribution Centre Type 1 Distribution Centre to Business

Business Type 2 Business to Business

Type 3 Business to Consumer Consumer Type 4 Cosumer to Consumer

Figure 2.3: A typology of Last Mile Delivery system

The last mile is the most inefficient and most costly part of the supply chain (Giuliano 2013; Rodrigue et al. 2013) and it poses significant urban planning problems (Boerkamps, et al., 2000). In addition, environmental pollution, safety concerns and infrastructure damage are associated with last mile delivery vehicles (Song et al., 2009). The dynamics associated with last mile delivery result in the un-predictability of last mile delivery within the city structure.

Overall, last mile delivery sits in the last hierarchy of the supply chain and city logistics systems. Its inefficiency and associated problems lead to high cost of this last leg of the supply chain process, hence, it is the weakest part of the entire process. Delays and inefficiency at this leg can be clearly perceived and costly to both supplier and the customer.

2.3 Perspectives on Last Mile Delivery

Last mile delivery as a part of city logistics has been theorised and modelled from a range of perspectives. These perspectives are discipline-driven and provide the broader basis for formulating frameworks for examining last mile logistics. These include the supply chain perspective (Gevaers et al., 2014), transaction cost perspective (Aized and Srai, 2014), spatial perspective (Ehmke 2012; Morganti, and Gonzalez-Feliu 2015) and the urban planning perspective (Aized and Srai 2014; Lindner 2011). These are discussed as below.

2.3.1 Supply Chain Perspective

Last mile delivery is a key component of the supply chain at the downstream end. In a globalised marketplace, SCM are increasingly becoming customer-oriented (Tan et al., 1999) as the end consumer drives demand in the supply chain. SCM therefore can be characterised as ‘demand chain management’ (Juttner et al., 2007 pp. 3). This is the reason why last mile delivery is indicated as the most important part of the supply chain process (Ewedairo et al., 2018). Supply chain process is not completed, until the product gets to the customer.

SCM is a management philosophy with integrated behaviours and processes, (Mares, 2010) with primary focus on how the logistics activities and operational decisions improve the last mile logistics performance (Vaaland and Owusu, 2012) as well as coordination and cooperation among the SCM operators from the first to the last leg of the activities (Schwartz and Carroll 2003; Xu et al. 2013). Overall, three major themes and six subthemes can be extracted from the existing literature.

The main themes are Activities, Benefits and Constituent components. As shown in Figure 2.4, material/physical, finances, services, and information flows; internal and external networks of relationships, as well as value creation, creation of efficiencies, customer satisfaction, constituents or components as the six sub-themes (Stock and Boyer, 2009). Last mile delivery as a service relates to the happenings at the final stage of the supply chain management. This represents the distribution of goods to end consumers. It also includes the external network (transport networks in this instance) through which the services are carried out.

Figure 2.4: Themes and subthemes of SCM (Source: Stock and Boyer, 2009)

Overall, last mile delivery falls within the distribution stage of the three traditional stages within the supply chain management. The three stages in the chain described as a system includes procurement, production and distribution (Thomas and Griffin, 1996).

2.3.2 Transaction Cost Perspective

The transaction cost perspective deals with operation costs and assets involved in the last mile delivery. Transaction Cost Economics (TSE) is a social theory developed by Coase (1937) and extended by Williamson (1977). The theory offers a set of normative rules for choosing an exchange relationship from alternative arrangements (Ghoshal and Moran 1996). It attempts to provide explanations for economic considerations by firms, as well as the advantage of transactions within and between enterprise and businesses including cost negotiations (Black et al., 2017).

The focus of TCE is on the cost required and the amount of effort for two persons to conduct businesses, otherwise known as “economic exchange or transaction” (Williamson 1977; Oliva and Watson, 2011). This can be related to delivery service between the warehouses or retailers, last mile delivery operator and the enduser in this instance. The outcome of the

effort and cost is influenced by assets specificity, uncertainty and governance mechanisms (Malhotra, 2003).

Williamson, (1977) identified three critical variables that characterise a transaction. These variables are (i) the uncertainty under which the transaction takes place, (ii) the level of transaction specific investment and (iii) the frequency that transaction occurs. Last mile delivery is clearly characterised by all three variables. Firstly, constraints of the environment create different levels of uncertainty for last mile delivery. Secondly, there are different levels of interests to be satisfied and thirdly, the frequency of last mile is so high that any resources invested in it are long term investment. All these can be associated with the high cost in last mile delivery.

The cost of last mile delivery become significant in the occurrence of transaction-specific investments and uncertainty (Williamson, 2008; Oliva and Watson, 2011). The problem of transactional uncertainty is associated with the decision environment in which a transaction occurs. While some investment can be controlled by last mile delivery operators, the controls of other investment are outside the last mile delivery operators.

Uncertainties in a transaction completion can come in many forms which can include the uncertainty created within a changing urban environment. Ewedairo et al., (2018) identifies the constraints of the built environments, transport network and planning constraints that can cause such significant uncertainty and delay in last mile delivery. The uncertainty associated with the complexity of compact city, increased vehicular and population densities in activity centres, and narrower road are factors that can account for uncertainties when goods are to be delivered within the CBD. This form of uncertainty creates problems of adoption, which, in turn, results in an increase of transaction costs for adaptation of new circumstances. The objective of the consumer and the retailers in last mile delivery is to minimise the cost associated with the uncertainty of their transaction (Lai et al., 2005). The uncertainty and asset uniqueness associated with the transaction determines the efficient mode of relationship governance (Walker and Weber 1984).

The use of transaction cost perspective affords clarifications for optimisation of last mile delivery associated costs. It emphasises the importance of the selection criteria in minimising transaction costs with different last mile delivery process (Sancha et al., 2016).

2.3.3 Spatial Perspective

Spatial perspective views last mile delivery as a location service (Stock and Boyer, 2009), which takes place within the constraints imposed by geographical space (Linder 2011). Spatial perspective focuses on understanding spatial relation and connectedness of last mile activities such as pickup and drop off points.

Understanding the impacts of landuse on transport and transport on land uses as relates to density, location accessibility, travel cost and time is therefore important in the discussion of the city live (Wegner and Fuerst, 1999). Classical theories such as the central place theory, concentric model, sector model and the multiple nuclei model provide a spatial perspective to capture the spatial structures of cities that influence the transportation systems.

The central place theory posits centrality with size and distribution of settlements. It attempts to establish the reasons why settlements are distributed the way they are. The range of goods influences the shape, size and distribution of cities/towns which can be predicted by three key principles i.e. marketing, transportation and administrative regions (Christaller, 1933; Brian et al, 1958; Getis and Getis, 1966). Concentric zone (Burgess, 1935) theory is based on human ecology, which argues that scarce urban land and competition for land create five concentric zones. A city centre, immediately surrounded by an area of deterioration, as zone of transition with the fifth and last zone is commuter zone. The sector model (Hoyt 1939) is considered as a revised form of the concentric zone model considering the emergence of series of sectors as the city grows from the CBD. To Harris and Ullman (1945), though the city may have stated with a CBD, other smaller CBD emerges on the outskirt of the city creating other nuclei in other parts of the city. This is the position of the multiple nuclei model. The city structures affect the way transportation systems are arranged, organised and interconnected. These spatial structures affect the requirements for last mile delivery shorter delivery time, and reliability as well as flexible delivery and management of associated delivery information.

Urban form and urban structure are intricately linked to the efficiency of LML. In terms of city logistics, the application of polycentric rather than monocentric city structure with appropriate transportation links often reduces long LMD travel times. LMD in a polycentric city with more decentralised economic functions and industrial activities generally require shorter LMD distances compared with a monocentric city. Retails models such as Hotelling

Location Model (Andre et al., 1987), Reilly Retail Gravitation Model (Reilly, 1953) and Gravity Model (Losch, 1954) can help design the spatial structure to optimise retail networks to minimise LMD. The identification of the least cost (Weber, 1909) or profit maximisation (Losch, 1954) location for optimal spatial structure of retail logistics enables delivery lead time and cost to be reduced.

Frost and Sulivian (2012) projected the future mega trends in last mile delivery challenges observing that by 2025, there will be 35 mega cities globally with requirements of 500 million deliveries required per person per day to cites and that demand delivery will become increased and more frequent, hence multi modal, low carbon footprint through utilisation of road and rail within the city will be required. Further studies have shown that by 2050 the world will 34% and 66%, representing one third and two third respectively (United Nations Department of Economic and Social Affairs 2012; 2014). This trend in the population growth and continued urbanisation will continue to impact the location, distribution/transportation, sorting and warehousing and ultimately impact the traditional last mile delivery (Hesse, 2002).

2.3.4 Urban Planning Perspective

Urban planning perspective views last mile logistics as a part of the landuse development and design as well as the infrastructure of the built environment. These infrastructure include necessary network to facilitate the process of movements of people, information and goods. Therefore the key objective of the urban perspective is to formulate plans which optimise the effectiveness of transportation infrastructure, services and landuse. The built environment is governed by landuse controls and the transport regulations. For instance, urban compactness is used as a mechanism for controlling and regulating urban sprawl by promoting a relatively high-density, mixed land-use city structure supported by a more efficient public transport system and increased opportunities for walking and cycling (Chhetri et al., 2013). Specifically, in Victoria, the urban system is governed by the National Transport Commission Regulations - Australian Road Rules - and the Victoria Planning Provision. The Victoria Planning Provision derives its legislative backing from the Victorian Planning and Environment Act 1987 (The Act).

What can be deduced from emergence of compactness is that last mile city logistics is lacking in what is considered in the discussion of city compactness. While higher degree of

accessibility is considered, this is often in terms of passenger traffic without any thinking on the servicing of the compact city created. In the creation of intensive use of urban land, roads are partitioned to create bicycle paths, dedicated public passenger bus lanes with more controls on last mile delivery vehicles, lowered speed limit etc. Resulting from these measures, Morris (2009), identifies that last mile delivery continues to cycle around as a result of lack of loading space.

Planning control in Victoria is one of the functions of Councils. The different Council’s Planning Schemes is the instruments of planning controls in these municipalities. In addition, Council also carried out controls through Councils Local Laws. Planners utilise various tools to control last mile delivery vehicle. These tools can be direct or indirect. The loading/unloading control and enforcement as direct while the planning zones impact last mile delivery indirectly. Different planning zones will impact last mile delivery in different ways. For example, areas of mixed use and high-density development will impact last mile more than other area with low density and reduced population density. The National Transport Commission (Road Transport Legislation) Regulations 2006 (Australian Road Rules) governs the transport systems/controls.

2.4 Urban Form and Last Mile Logistics

Urban form refers to the shape, size, density, land use and arrangement of the built environment (Anderson et al., 1996). These features are both physical and non-physical that characterise the city (Alho and Abreu e Silva, 2015). These features make each city unique and LMD is one such feature of the uniqueness of each city.

Each city demonstrates different forms, a combination of forms and elements, or an agglomeration of perspective elements (Lynch, 1981), which Jenks et al. (2008), broadly categorised into five broad and interrelated elements. These elements include: density, housing and building types, layouts and landuse as well as transport infrastructure. Lynch (1981) categorised urban form on the basis of street patterns, block size and form as well as street design, lot shape and layout. Cities are also characterised by the spatial imprint of the urban transport system, adjacent physical infrastructures and socio-economic activities (Rodrigues et al., 2015). Pressman (1985) and Minnery (1992), after Newton (1968) identify five archetypal urban geometrics to include dispersed city, compact city, edge city, corridor

city and fringe city. These five archetypal geometrics can be differentiated by the level of centralisation or clustering (Rodrigue et al., 2013). Urban transport system in the form of spatial imprint and spatial interactions takes pre-eminence in any of these archetypal urban geometrics (Rodrigues et al., 2015).

Broadly, the urban transport system is characterised by three elements of infrastructure, modes and users that shape urban form (Doel and Hubbard 2003; Graham and Marvin 2002). Different socio economic, geographical characteristics as well as circulation patterns of passengers’ movement and freight drive spatial imprint and spatial interactions. This is graphically illustrated in Figure 2.5.

Figure 2.5: Transportation, Urban Form and Spatial Structure (Source: Rodrigue et al., 2013)

Research into the relationships relating last mile delivery to urban form is not yet well entrenched in the extant literature. However, studies that examined the relationship between passenger transport and urban form are widely prevalent across a range of different geographical scales (Stead and Marshall, 2001). The recent studies on last mile and urban forms includes Allen et al. (2012) and Alho and Silva (2015). While the work of Allen et al., (2012) determines whether particular urban designs and city layouts limit traffic volume; Alho and Silva, (2015) assesses last mile delivery generation based on the distribution of population within Lisbon, Portugal The findings of McKinnon, (2010 pp. 21) submits that last mile delivery is affected by “geographical location, land use patterns, accessibility to local transport infrastructure/networks, availability of parking facilities, and road network type” that can be considered as impedance. This impedance thus in turn influences the efficiency of last mile delivery (i.e. the ratio of tonne-miles to vehicle miles) within the city.

Of relevance to this, transport growth can affect the particular forms of urban development (Alho and Silva, 2015).

The body of the research on passenger transport and built environment ranges from the regional strategic planning level to local planning issues at neighbourhood level with all levels of government showing interest (Stead and Marshall, 2001). However, passenger journeys and last mile delivery are very different in nature. While passenger journeys are decided on (in terms of origin and destination, mode, timing and frequency of travel) by individuals for a multitude of different reasons, last mile journeys are a final journey, a single form of transportation that aims at delivery of goods. Decisions about the nature of the last mile are made by private companies to the customers they do business with. Hence coherent urban freight transport policies have not been developed to the same extent they have for passenger transport (Cherret et al., 2012).

Many urban authorities have begun to focus on the efficiency and sustainability of freight transport due to its economic importance over the last decade. This has led to efforts to develop freight transport strategies and plans in some cities using a combination of measures outlined by Stathopoulos et al. (2012), as well as research projects, trials and operational schemes. These measures, as highlighted by Cherrett et al. (2012, pp. 78), “include the implementation of urban consolidation centres in French, Italian, Dutch and British cities; the establishment of Freight Quality Partnerships in many British cities; the development of quieter freight operations to facilitate out-of-hours deliveries in Dutch cities; the variable use of road space by time of day in Barcelona; the use of electrically-assisted tricycles for parcel deliveries in central urban areas such as London, Paris and Brussels; and the use of locker banks and collection points in German, French and Belgian cities”.

Transportation planning process is much dependent on individual planners and policy makers, since they may have different ways of interpreting concepts such as sustainable development (Quak, 2011). If local authorities are to achieve their objective of a sustainable city transportation system, they need to be aware of the activities involved as well as their potential impacts.

In the study of conventional transport network analysis methods, Lowe (1975), uses the planar representations for reduction of intricate transport system into nodes and links. While nodes denote junctions, links are roads or streets connecting the nodes. Planar graph refers to

a set of two-dimensional systematically arranged points and lines. Points (nodes) and lines (links) are therefore the basis of spatial analysis (Bowen 2012; O’Kelly 1998).

Points and lines in terms of transport networks are creation of the city’s growth and landscape as well as location, topography, climate, culture and other important elements of the city (Thomson, 1977). Planar graphs of the patterns of the links reveal the uniqueness of these important city elements. These important street pattern as presented by different authors are presented in Table 2.3 and Figure 2.6 as described by Walker and Manson (2014) after Marshall, (2005). The table above shows the inconsistency in the use of terms/concepts to define and describe these patterns. However, it is possible to discern a much smaller of recurring subsets of network pattern. Walker and Manson, (2014). Brindle (1996) argued for adoption of two ‘fundamental’ urban street layouts, the grid and the tributary street network, to promote transportation efficiency.

Table 2. 3 : Descriptive terms applied to urban street network

Author Descriptive terms Unwin (1920) Irregular Regular Rectilinear Circular Diagonal Radiating lines Moholy-Nagy (1969) Geomorphic Concentric Orthogonal-connective Orthogonal-modular Clustered Lynch (1981) Star (radial) Satellite cities Linear city Rectangular grid cities Baroque axial network The lacework Other grid (parallel, triangular, hexagonal) The ‘inward’ city (medieval, Islamic) The nested city Current imaginings (megaform, bubble floating, underground, etc.) Satoh (1998) Warped grid Radial Horseback Whirlpool Unique structures Frey (1999) The core city The star city The satellite city The galaxy of settlements The linear city The polycentric net (Source: Walker and Manson, 2014)

In last mile delivery, transport network reflected in the grid street network will enhance efficient delivery compared to the linear, tributary and radial street networks.

Figure 2.6: Urban street pattern archetypes (Source: Walker and Manson, 2014)

2.5 Last Mile Delivery Impedance

Last mile delivery impedance is a friction of space (Hesse and Rodrigue, 2004) and time constraint (Sivakumar and Bhat, 2007). Hesse and Rodrigue (2004) considered the multidimensional aspect of the concept impedance as relates to transport costs, organisation of the supply chain and built and regulatory environment as well as location of where the last mile transaction takes place. All these elements define the concept of last mile delivery impedance (Hesse and Rodrigue, 2004).

Of all the elements, the built and regulatory environment constitutes an important constraint which directly results in impedance through the planning and transport systems. This built and regulatory environment is the ‘material space’ (Hesse and Rodrigue, 2004) where the last mile delivery movements is embedded. These are the transport infrastructure like road, railway, warehouses, and terminals necessary for an efficient last mile delivery of goods to end consumers. These external determinants of last mile delivery play a vital part because of the critical pressure exerted on last mile delivery during the peak hour use of transportation infrastructure in large cities, especially inner city suburbs and main arterial roads and freeways. Hesse and Rodrigue (2004), identified infrastructure road bottlenecks and congestion, urban density and urban adjustments from compactness as impedance measures to last mile delivery.

In this thesis, the concept of impedance is defined as the amount of resistance required to traverse through a route on a road network from the pick-up to the delivery point. High impedance on a road network deters movements of goods in and out of urban areas as well as goods travelling within the city (Ewedairo et al., 2018).

In simple cases, network designs represents visual illustration that allow differentiation of one network from another. However, in a more complex instances, the visual complexity makes the task of reliable and objective identification of the differences difficult (Walker and Manson, 2014).

Rodrigue et al., (2013) differentiates between locational attractiveness and emissiveness that can be applied to last mile delivery. Accessibility has to do with attractiveness and emissiveness of locations. While attractiveness relates the measure of the capacity of a location to be reached, emissiveness is the capacity to reach different locations (Rodrigue et al., 2013). Locations of high accessibility are likely to have more people travel to them or through them than locations of low accessibility because such location can be more easily reached and are also more integrated and connected to than other areas within a city (Richards-Rissetto and Landau, 2014). However, a one-to-one relationship between accessibility and integration is unlikely to be established within the city. Accordingly, a place that is easily accessible will be more integrated with the economy of the city (Hillier et al. 1993; Hillier 1996).

The volume of last mile delivery to or from a location or that passes through a location therefore strongly correlates with the levels of integration and connectivity with the city (Bafna 2003; Hillier 1996; Ratti 2004) and delays within the network. The ease of reaching the location of delivery from pick up is a function of the level of impedance posed by the amount of resistance within the transport network.

2.6 Summary

This chapter described the key concepts used to reflect last mile logistics and the theoretical perspectives applied to examine the position of last mile within supply chain management and city logistics. The chapter provided key definitions of city logistics and last mile. Last mile is described as part of city logistics and the final leg within the supply chains system, the most costly and inefficient of the three parts of the supply chain.

Four different theoretical perspectives to understand last mile delivery within the supply chain process are discussed. These comprise the supply chain, transaction cost, spatial and planning perspectives. The spatial and urban planning perspectives inform the discussion on

urban form with last mile delivery impedance and the transport network, spatial accessibility and impedance discussed in section 2.4 and 2.5 of this chapter.

Chapter 3 proceeds to conceptualise and present theoretical framework for building the future of last mile delivery scenario.

CHAPTER THREE

CONCEPTUAL FRAMEWORK FOR BUILDING THE FUTURE OF LAST MILE DELIVERY

3.1 Introduction

In the preceding chapter, the concept of last mile delivery was discussed and different theoretical perspectives to examine its forms, functions and performance were presented. This chapter develops a theoretical framework to systematically establish the linkages between key constraints and last mile delivery impedance and build last mile delivery scenarios. Creating scenarios to understand which situation will eventuate in the future of last mile requires intuitive thinking of the key stakeholders involved in last mile delivery.

This chapter is organised into six sections. The theoretical framework upon which the thesis is hinged in section 3.2. The section conceptualises the interrelationship between last mile and the urban environment and proceeds to identify constraints that impedes last mile delivery with the development of the conceptual framework for this study. The Theory of Constraints and the embedded thinking process are discussed in section 3.3. Section 3.4 presents Scenario Thinking as a tool to build plausible last mile delivery scenarios and the methodology of the research for projecting last mile delivery scenario and the resulted likely outcomes. Identified last mile delivery constraints are discussed in section 3.5. Section 3.6 presents the summary of the chapter.

3.2 A Theoretical Framework for Building LMD Scenarios

Last mile delivery poses significant urban planning problems and logistics challenges (Boerkamps et al., 2000). The dynamics of last mile delivery will continue to change given the rapid increase in e-commerce, accelerated rate of globalisation, technological improvements and digital disruptions, and changes in business practices including reduction in brick and mortar (warehousing spaces) through e-commerce (Lee et al 2001; Hubner et al, 2016). Furthermore, changes in consumption patterns (Hubacek et al, 2001), transformation

of built urban environment (Agunbiade, 2012) and growing interest of different stakeholders in city logistics and greening of urban freight (Ballantyne et al., 2013) impart further uncertainty in future of last mile logistics. City administrators and policy makers are redesigning inner city and arterial growth corridors based on the compact city model to solve increasing urban population and urban sprawl in outer growth areas. These urban transformations in spatial structures and urban functions put intense pressure on last mile logistics. Thus, the need to know the future of last mile logistics in large cities to cope with the rising demand for home delivery of goods within a compact city model, setting ground for the need to build future scenarios.

Of implication of this assumption, the future is uncertain, and no policy will be appropriate for all time, hence as the urban environment changes, last mile delivery will also change and will necessarily adapt. As the transport networks are part of the established built environment, planning control impedance on last mile can be considered as an increamentalist approach, which (Fainstein & Fainstein, 1996) viewed it as a non-planning approach, based on laissez-faire premises (Alexander, 1979; Fainstein & Fainstein, 1996). However, even though incrementalism may be regarded as the opposite of planning, it has gained much attention within planning theory, as “it produces the fruits of planning in its results” (Fainstein & Fainstein, 1996, pp. 272).

The development of theoretical framework for building last mile delivery scenario is hinged on the two principles of Theory of Constraints. The first principle of logistics paradigm and second principle of Theory of Constraints (Goldratt and Fox 1986; Rahman 1998), the thinking process, when coupled with Scenario Thinking, provides a comprehensive framework to identify and measure last mile impedance and build last mile delivery scenarios (Figure 3.1). In this thesis, an integrated approach is adopted to combine the applicable parts of the two principles and embed them in Scenario Thinking framework.

What to change?

Theory of Constraints

What to change to? Scenario Thinking

Identification of System Constraints Thinking Process

How to cause the change?

Figure 3.1: Key questions linking the Theory of Constraints and Scenario Thinking

Figure 3.2 represents the overall theoretical framework linking the Theory of Constraints to Scenario Thinking. The dynamics from the emergence of compact city are reflected through the built and regulatory environment represented by infrastructure and population, and policies that shapes the built environment respectively. This was carried out using the spatial measures of the built environment, the planning and transport systems as constraints through the first step of logistics paradigm and thinking process (the principles of TOC). The constraints are represented as last mile delivery impedance as Current referring to ‘what to change?’ While Future relates to ‘what to change to?’ and ‘how to cause change?’ questions of the thinking process. The different scenarios and last mile delivery strategies assist in answering these questions.

Strategy5

Strategy4

Strategy3

Strategy2

Strategy1

Last mile delivery Strategies mile Last

controls

Scenario 4 Scenario

Scenario 3 Scenario

Scenario 2 Scenario

Scenario 1 Scenario

Future

Delivery

Last Mile Last

Impedance

ights

Toll

Zones

Centre

Industrial

Residential

Speed limit Speed

Intersection

Commercial

Tram Routes Tram

Traffic count Traffic

Traffic Traffic

Activity Centre Activity

Number of lanes of Number

Population Density Population

Railway Boom gate gate Boom Railway

Proximity to Activity to Proximity

Scenario Scenario

Thinking

Theory of of Theory

Constraint and and Constraint

Built

System

Planning Planning

Transport Transport

Environment

Framework for linking last mile delivery with urban structure and planning planning and urban with last structure linking mile delivery for Framework

Theoretical

Regulatory

Population) Population)

Planning and Planning

Environment Environment

BuiltEnvironment)

Built Environment BuiltEnvironment

(Infrastructure and (Infrastructure

(Policiesthatshape the

Transport controls Transport

Figure 3. Figure

Compact City Compact Current

3.3 Theory of Constraints (TOC)

In the 1990’s, Goldratt developed the Theory of Constraints to examine how firms can achieve their goals or improve performance. The Theory of Constraints has been applied in the fields of social and management sciences (Blackmore, 2001; Tsou, 2013) including Finance, Project Management, Marketing and Supply Chain management (Goldratt and Cox 1992; Simatupang, 1997; Rahman, 2002; Simatupang et al., 2004; Tsou, 2013). The Theory of Constraints describes what is holding the system back and what to do about it (Goldratt, 1990). Constraints are `anything that limits the performance of a system relative to its goal’ (Goldratt 1990). The constraint is any element or factor that confines the system from doing more of what it was intended to achieve (Goldratt and Cox, 1992).

The philosophical principles that underpinned TOC are two separate yet interrelated components (Simatupang et al, 2004). These include: Logistics Paradigm and Thinking Process. The thinking process is termed logical thinking after Simatupang, 1997; Cox and Spencer, 1998; Rahman, 2002; Simatupang, 2004). These two are the underpinning working principles of TOC which are discussed as following.

3.3.1 The First Principle: The Logistics Paradigm

The logistics paradigm (LP) refers to the scheduling procedure and the buffer management information system. This principle involves five major steps to reflect on-going improvement of the system. These include: (i) identification of system constraints (ii) exploitation of system constraints, (iii) subordination of all resources, (iv) elevation of constraint and (v) overcoming inertia (Cox and Spencer 1998; Rahman 1998, 2002; Dettmer 2007). Goldratt (1990, pp.135) linked these five steps to the process of continuous improvement (Figure 3.3).

Overcoming inertia Elevation of constraint Subordination of Exploitation of everything to the system constraints exploitation decision Identification of system constraints

Figure 3.3: Logistics paradigm

Identification of system constraints is to enumerate and recognise the bottleneck in a last mile logistics system, which could be either positive or negative. The constraints could be physical or material (Rahman 2000) and in last mile delivery, could be related to the human dimension. Exploitation of system constraints relates to a decision process which assess the optimal methods of reducing the impedance imposed by the identified constraints. This is the step where the constraints are broken down, managed (Blackstone, 2001) or eliminated (Rahman, 2002). Subordination relates to subservience of all constraints to the exploitation decision. It applies to adjusting all resources to support the effectiveness of one prioritised constraint (Rahman, 2002). This is to reduce the limit impact of the constraint (Goldratt, 1990) on the system. Elevating system constraints purport at improving positive constraints until it achieves its best performance. Elevating and concentrating on improvement of positive constraints will bring the system to a stage where it will no longer limit the performance of the system. Overcoming inertia – requires going back to identification of constrains and repeating the four steps if in any step a constraint is broken (Rahman 2002). This is because, a new constraint will require a new strategy.

Out of the five steps of the logistics paradigm principle listed above, the identification of constraints step is the main one that links the theory to this research. It is this step that connects the thinking process and Scenario Thinking methods used that underpin this research. The other four steps of the logistic principle is applicable to production line, marketing and organisations where constraints can be controlled. The uncertainty of the changing built environment and the human dimensions of the constraints does not allow opportunity to control the constraints identified in the research. In addition, the constraints of the built environment, planning and transport network on last mile can be positive and at the same time negative.

The broader objectives of last mile delivery are efficiency, cost effectiveness and sustainability. To optimise these objectives, constraints are usually considered to be negative (Sivakumar and Bhat, 2007). Constraints are conceived as impedance (Hesse & Rodrigue, 2004), which need to be reduced or totally eliminated (Blackmore, 2001; Tsou, 2013). In addition, operational constraints characteristics of the urban environment such as planning and transport network limit last mile delivery efficiency and cost effectiveness.

Rather than being rigidly positive or negative (Rand, 2000), constraints related to the built and regulatory environment in the last mile delivery swings between being positive and

negative (Ewedairo et al., 2018). There is however no theoretical framework that establishes the linkages between the last mile delivery and built and regulatory controls. The Theory of Constraints provides a base for evaluating the impact of urban form, spatial structure and regulatory processes that shape the scale, intensity and patterns of last mile delivery in an urban setting. This is in line with the submission of Simatupang, (1997) who noted constraints to be physical, behavioral, and policy driven (regulation and guideline).

Specifically, Sivakumar and Bhat, (2007) relate constraints to the levels of impedance. The concept of impedance is vital the requirements of many economic and social geographical review condition in last mile delivery paradigm (Button, 1993; Bowersox et al., 2000). Ewedairo et al., (2018 pp. 111) define last-mile delivery impedance (LMDI) as the amount of road network resistance imposed on LMD that limit the efficiency of the delivery. These impedance hold back the performance of last mile delivery resulting in being the most costly and inefficient part of supply chain. Last-mile delivery then represents the weakest point of the supply chain which constraint the objective of the supply chain system.

3.3.2 The Second Principle: The Thinking Process (TP)

Thinking process is a way of combining ideas to form a logical and coherent reasoning. The thinking process enhance the system to adapt to change. In order to manage constraints, Goldratt (1990) has identified three important decisions which need to be considered as a part of thinking process. These include deciding what to change by pin pointing the problem, what to change to construct the solutions and how to cause the change through deploying appropriate resources (Goldratt 1990). The underlying aim behind the thinking process is to identify the core problems, develop practical solutions and implement the solutions (Rahman 2002).

Five cause and effect tools are developed in the thinking process to address these questions on change (Rahman, 2002). These tools include Current Reality Tree (CRT), Evaporative Cloud (EC), Future Reality Tree (FRT), Prerequisite Tree and Transition tree. The roles of each of these tools are presented in Table 3.1.

Table 3.1: Thinking process tools

Thinking Process Purpose Thinking Process questions Tools What to change? Identify core problems Current reality tree What to change to? Develop simple, practical solutions Evaporative cloud How to cause the change? Implement solutions Future reality tree Prerequisite tree Transition tree (Source: Rahman, 2002 pp. 812)

Current Reality Tree (CRT) is the first tool of the thinking process that examines the cause and effect relationship. It seeks to analyse the problem of a system, core problem or underlying cause (Blackstone, 2000). This tool assists in the identification of “what to change” in a system. It is a logical structure that depicts the reality in its current state (Rahman, 2002). Understanding the root cause of a problem will assist in focusing on the problem rather than the symptoms. Evaporative Cloud (EC) expresses the core of conflict and surfaces hidden assumptions. This tool is often used in conflict management (Gupta et. al., (2011). It is a tool used to identify and display underlying elements and assumptions of a conflict (Goldratt, 1990; Dettmer, 1997) and develop plausible practical solutions that invalidate one or more of the assumptions (Rahman, 2002). Future Reality Tree (FRT) is the third step in thinking process. This is a tool used to anticipate the future or plan on how to implement a major solution for an improved future. It turns a future that is undesirable to one that is desirable through actions. Prerequisite Tree is a logic based process to achieve an objective. The extent of achievement at each of the stages is mapped and measured. It serves as a preparatory process of implementation of an action. Transition tree is the last tool of the thinking process, which details the action required towards the implementation of a plan outlined on a Prerequisite Tree.

The application of TP to this study hinges on its ability to strategically consider major future changes through logical thinking. As much as novel these tools are (see Dettmer, 1997 and Rahman 2002), there application to the last mile delivery problem however is not simple but rather complex. This is due to the complexity of last mile logistics, which are driven by internal operational and external environment. Last mile logistics is not like the business operations within a firm or a supply network where constraints can be easily managed and controlled through the implementation of an action. The level of uncertainty of the external environment is much higher in city logistics provisions. Scenario Thinking is another

approach which if integrated with the Theory of Constraints would provide a broader framework to capture the complexity of last mile logistics. The Scenario Thinking is also similar to FRT that considered the desirable future. All these TP tools are to identify the core problems, search for plausible solutions, plan changes and implement the changes. Further, they seek to resolve the issue of how to do it and how to change (Dettmer 1997).

Overall, Scenario Thinking specifically relates to the first step of the first principle of TOC - identification of constraints and answer the question of what to change by pin pointing the problem, what to change to, construct the solutions, and how to cause the change through deploying appropriate resources (Goldratt, 1990). In this research, key constraints that impede last mile delivery are what to change, the best scenario with desirable outcomes (i.e. efficiency, sustainability) is what to change to and the strategies to drive how to cause the required change.

Identification of these constraints, which is step one in the first principle of TOC represents what to change in the thinking process – the second principle of TOC. What to change will be the constraints of the built environment that impede last mile delivery. Breaking the constraint down and managing or eliminating the constraint (exploitation). Adjusting these constraints through subordinating the negative ones and elevating the positive constraints through the current reality tree is therefore important. The future reality tree through the use of Scenario Thinking methods will assist analysing what to change to and how to cause the change.

The Scenario Thinking can be employed to analyse current situation (what to change), trends and building scenarios (what to change to and how to cause change). Analysis of current issues includes constraints identification can be conducted through Scenario Thinking. The methodology section discussed the five steps approach utilised in the research.

3.4 Scenario Thinking Method

Scenario is considered to have its origin in the field of theatre and movies to describe play’s outline. In ordinary term, two meanings or definitions come to mind when the word scenario is mentioned. Firstly, as an outline of a film, or a play which details the different scene and secondly as sequence of imagined future events (Wack, 1995). The second meaning has been used in the military, risk management policy advice and planning (van Notten, 2005). This

description of scenario is adopted by Intergovernmental Panel on Climate Change (IPCC), (2008) and it is how scenario is considered in this research.

The term scenario has been applied in many fields of studies. IPCC (2008) described a scenario as a coherent, internally consistent and description of a plausible future state. Scenario is not a forecast; hence each scenario is an alternative appearance of how the future can eventuate. Aligned with this definition, Mahmoud et al. (2009, pp. 798) defined scenarios as “possible future states of the world that represent alternative plausible conditions under different assumptions”. Scenario is therefore the description of possible happenings in the future given certain circumstances of events, presenting the image of the future (Schwartz, 2000). What is not certain is how the “circumstances of or several event(s)” will unfold in the future. This creates uncertainty.

The uncertainty about the future is what has resulted into the quest for human to be in the business of looking for how the future will unfold. Scenarios generally aims at sensitive issues considered important to stakeholders. It therefore allows decision-makers to anticipate coming change and timely responsively prepare for the change (Schwartz, 2000; Santelmann et al., 2001; Steinitz et al. 2003). The use of scenarios has the ability to contest orthodox thinking and previous accepted norms for the future (Kahn & Wiener, 1967) with the objective of producing a small number of scenarios with plausible explanations of system factors that can be possibly different in each scenario (Mahmoud et al., 2009).

The word planning (Schoemaker, 1991), method (Wack 1995; Sarpong 2011), approach (Jungermann & Thuring 1987; Burt 2006; Kosow & Gabner 2008), thinking (Mackay and McKiernan 2006; Wright & Cairns, 2011) are common suffixes associated with scenarios. The suffixes are variously used depending on the context at hand. This research applies the syntax “Scenario Thinking” after Wright & Cairns, 2011 and Sarpong (2011). Sarpong (2011), applied the term ‘Scenario Thinking’ from Mackay and McKiernan (2006). This is due to the fact that outcome of the exercise is from brainstorming and intuitive reasoning of participants.

Scenario Thinking approach provides an opportunity to bridge the void and the uncertainty that cloud the understanding of the future in the present time. Its emergence therefore can be traced to the application of ideas of foresight studies (Sarpong, 2011; Burt, 2006). When linked to TOC, it can be described as a subset of future reality tree (FRT). The approach of

Scenario Thinking to understand and analyse problems that apparently intractable is unique. Especially where there are ‘critical uncertainties’ that span a range of subject areas or boundaries of many discipline (Wright & Cairns, 2011) like last mile delivery. Traditionally, last mile delivery span through manufacturing, marketing and planning as well as transportation.

Its application allows discovering of innovative foresight where there are enhanced continuous change, countless intricacy and higher uncertainty. All these are drawn through internal and external sourced methods and competences towards future insight (Wack, 1995). Hence, Scenario Thinking can be used in conjunction with qualitative or quantitative information and data, and can provide input to other methods. It can also be used by individuals, groups with different social science backgrounds (Jungermann & Thuring, 1987).

3.4.1 Scenario Thinking - Schools of Thought

From palm readers, star gazing or guessing through to scientific interpretation and probing, human has been exploring, discovering and projecting the future and developing ways and methods to solve the challenges that might come with it (Bishop Et. al., 2007). More recent endeavours have logically thought of and analysed possible situations to rationally anticipate the future using both quantitatively (predictive) and qualitatively (descriptive) methods.

There are three basic schools of thought which were identified in Scenario Thinking research (Schnaars 1987; Schmenner 2001; Postma & Liebl 2005; Ramirez, et al., 2015). These schools of thought according to Boyonas et al., (2016) include:

 the Intuitive logic,  the probability modified trend and  the French-Origin La Prospective

The intuitive method, also identified as the Anglo-American school of Scenario Thinking. This method was introduced by Kahn in the US in the 1960’s. It is a “soft” foresight technique (Foresight, 2011) that includes employment of eclectic technique of thinking outside the box or brainstorming (Hollins, 1999). The Anglo-American Scenario Thinking method became popular through its successful application in the Royal Dutch-Shell Corporation (Wright and Cairns, 2011).

The second school of thought is identified as the Probabilistic Modified Trend that includes Trend-Impact Analysis (TIA) and Cross Impact Analysis (CIA). Scenarios based on historical data is the focus of the trend impact analysis, while the cross impact analysis introduces causal and correlational cross impact relationship in the scenarios (Boyonas et al., 2016). The French-Origin La Prospective School utilises statistical computer based approach as well as mathematical methods for scenario creation (Bradfield et al., 2005). The elements of each of the school of thought in Scenario Thinking are presented in Table 3.2.

Table 3.2: Summary of basic approaches in Scenario Thinking

Intuitive Logics Probabilistic Modified French-Origin La Trend Prospective School 1 Define the problem Collect historical data on Analyse problem and define topic system under examination 2 Collect all Use algorithm to select curve- X-ray the institution to be available fitting historical data examined (know how, information about product lines etc.) the future 3 Identify patterns Generate ‘surprise-free’ Identify key variables of the and trends future trends institution and the environment 4 Establish Scenarios Develop list of future events Understand dynamics of which could cause deviations institution and environment from the extrapolated trend (Strength and weaknesses etc.) 5 Use expert judgement to Reduce uncertainties (e.g. identify probability of event interviews with experts) and occurrence draw out most likely 6 Determine conditional Check on coherence probabilities of future events through cross impact calculations 7 Produce adjusted Assess strategic options extrapolations 8 Make strategic choice 9 Complete plan of action (Source: Ghaffar et al., 2005)

3.4.2 Methods and Techniques to Building Scenarios

There are four broad approaches and methods associated with these three school of thoughts (OECD, 2002). These include The Delphi method, Horizon Scanning, Trend Impact Analysis and Scenario Thinking.

Trend Analysis is the use of various techniques to analyse and project trends using historical data. Key steps in trend analysis are spotting an emerging trend, and examining what is causing the trend as well as other forces that may affect the trend through trend extrapolation (Hirsch et al., 1982). The core assumption of trend extrapolation is that changes in the future will continue to occur same way they have changed in the past. However, the change will stop at a point and level up (Guillemette and Turner, 2018). The stoppage of this change will therefore limit the extent of the trend analysis. Cyclical Pattern Analysis is used to tackle those changes which are recurrent and cyclical events which can be predicted in the future. Waves of events from recession to depression to revival-prosperity are evaluated and parameter estimated to project future trends. Environmental scanning is another technique that refers to the process of scanning prevailing conditions to identify emerging issues. The changing patterns are identified, changing trends and potential developments are evaluated and monitored, and then future patterns and their likely impacts are forecasted (Hirsch et al, 1982; Guillemette and Turner 2018).

Backcasting method is closely linked to the concept of scenario. Rather than the considering what futures are likely to occur, it considered desirable futures that can be created. In backcasting, the desirable future is imagined and works backward to determine required policy measures to achieve such a future (Dreborg, 1996). The six steps involved in backcasting includes: determination of objective, specification of goals and constraints, description of the present system, specification of the exogenous variables, undertaking of scenario analysis, and lastly, undertake important impact analysis. Backcasting aims at the generation of alternative descriptions of the future with a thorough analysis of their feasibility and consequences (Robinson et. al., 2011).

The usefulness of Delphi, horizon scanning, trend impact analysis and scenarios are highlighted by Guillemette and Turner (2018) in the study of the future. The Delphi method involves the gathering information through a series of questionnaires/interviews of selected experts and brainstorming over time to collect and iteratively confirm the deliberation and judgement on the issue. It gives opportunities for participants through the series of questionnaires to change and adjust their statements as may be required. This eventually results into reaching consensus (Regan et al., (2006). The problem of nonresponse to questionnaire is the bane of this method. This method is different from horizon scanning which detect early signs through a systematic examination of potential prospects and risks

(Guillemette and Turner, 2018) and exploration of new and unexpected issues. Horizon scanning can be a way to assess trends to incorporate in the process of scenario development. It supports qualitative and quantitative analysis and can as well be exploratory and sport unexpected future. Trend impact analysis takes into account unprecedented future events based on the present trend of events cumulating on how it will impact the future based on historical data (Dreborg 1996; Regan et al. 2006) through qualitative and quantitative methods.

Overall, the strength of scenario methods (which Herman Kahn considered to be the first person to relate scenario to planning, Schwartz, 1991 pp 7) over all other methods is presented in Table 3.3 as presented by Guillemette and Turner (2018). The strength of Scenario Thinking over all other methods lies in its ability to be applicable to different methodological approaches including quantitative, qualitative, normative and exploratory. Scenario Thinking is an approach to futurology (Kosow and Gabner, 2008). It can be distinguished generally by its orientation, whether normative or exploratory. Normative scenarios aim towards charting the desirable future to be achieved or avoided (IPCC, 2003; Milestad et al, 2014), while exploratory scenarios aim at identifying possible development regardless of their desirability (Milestad et al., 2014).

The process of engagement, its robustness and ability to spot unexpected (Guillemette and Turner, 2018) makes its stand out. Also, Scenario Thinking has an advantage of enhancing collective understanding and allows a democratic process where all viewpoints are considered equally and all ideas discussed through logical and intuitive thinking (Simatupang, 2004). Furthermore, it has been noted that planning with forecasting has not always been successful (Mahmoud et al., 2009). Scenario Thinking goes beyond forecasting as it determines the unlikely futures as opposed to likely futures allowing controllable assumptions about the future (Mason, 1998) through participation.

Participatory methods involve policy makers, stakeholders or a lay person playing active part in the scenario exercise. The advantage of this method is in its complementary capability to use expert knowledge that guarantees the sufficient richness (Anastasi, 1997). Participatory methods involve mutual learning method (Robinson et al., 2011) through stakeholder’s participation. These involvement improves the discussion through multiplicity of participants’ perspectives, skills and competences (ECDGXI, 1996 and Rotmans and Van

Asselt, 2000). This thesis applied the qualitative story based, participatory approach using five steps procedures.

Table 3.3: Approaches in the study of the future

Quantitative Qualitative Normative Exploratory Engagement Robustness Spotting Unexpected Scenario X X X X X X X Method Delphi Method X X X X X Horizon X X X Scanning Trend Analysis X X X (Sources: Foresight Toolkit U.K. and the AC/UNU Millennium Project)

Before understanding the future, it is necessary to understand the past, consider the present and use the knowledge to formulate the future. Sarpong (2011) noted that there is a need to have what he termed hindsight, foresight and prospecting in Scenario Thinking. Hindsight represents the past, while the foresight represents the future. Prospecting signifies development of insight into the present through peripheral vision. The methodologies to study these three perspectives of Scenario Thinking are classified under the general rubric of counterfactual analysis, peripheral vision and scenario analysis (Sarpong and Maclean 2011; Sarpong et al. 2013). The relationship between these three hindsight, prospecting and foresight in Scenario Thinking is presented in Figure 3.4 after Sarpong (2011).

Prospecting Peripheral Vision Foresight Hindsight

Counterfactual Scenario Analysis Analysis

Scenario Thinking

Figure 3.4: Embedded methodologies in Scenario Thinking (Source: Sarpong, 2011)

3.4.3 Time Horizon in Scenario Thinking

Scenarios thinking does not provide the answer to a problem or tends to predict the future, but the plausibility and the possibility of the future (Wright & Cairns, 2011) considering a range of time. It therefore provides a means of better understanding the complexity and ambiguity of the present (Kosow & Gabner, 2008) which offers an array of future possibilities for testing the current plans. In addition, develop and appraise new options as well as make better informed and considers the degree of possibility and plausibility.

A major focus is on how the future might evolve from today’s point in time to the horizon year of the scenario. The time scale for any scenario thinking is a function of issues under consideration and changes within that phenomenon and context dependent (Miller, 2005; van Notten, 2015). Van Notten, (2005) noted that in the fashion industry, 10 years is considered as a long-time horizon, while it is considered a short-time horizon in environment or business disciplines. The fashion world is in a constant change, while the effect of weather and climate over a longer period will impact any Scenario Thinking on the environment. Hence, scenario consideration can be short or long term.

While time is not a typology in its own right (van Notten, 2000), it is however relevant to establish a focus of direction which Scenario Thinking aims to address. The time horizon must not be too long or too short, it needs to be enough to create the probable and reflect its impact on the issues under discussion (Lindgren and Bandhold, 2003). The planning scheme affecting each Council in Victoria is expected to be reviewed periodically with a time frame of between 3 – 10-year periods. Therefore a 10-year time frame is considered appropriate for the Scenario Thinking horizon.

3.5 Identifying Last Mile Delivery Constraints

The urban environment is often represented by its forms, patterns and population. These includes both the tangibles (physical attributes) and the intangibles (air and water) components. The interaction between the urban environment represented by the built and regulatory environments are discussed in the previous section. The interaction reflects the current segment of the theoretical framework (Figure 3.2), using the measures of built environment, planning and transport systems derived from Scenario Thinking workshop and

literature. This section discussed the identified attributes of the built environment as well as the planning and transport systems as constraints to last mile delivery.

These constraints are discussed and evaluated on the basis of the current body of knowledge that highlights the linkages between compact city attributes as well as built and regulatory environment and city logistics. The Victoria Planning Provisions and the Australian Road Rules (Bowyer, 1988; Eppel et al., 2001; Barceló, et al., 2005; Blair et al., 2013; Kawamura, 2014) will be the key foundation that helps in the development of the theoretical framework to drive the analysis.

Historically, some of the policies adopted by local authorities which were implemented via urban planning and transport controls have to potential to directly or indirectly affect the efficiency and effectiveness of last mile delivery provision. As show in Figure 3.5, the interplay between last mile delivery, the built and regulatory environment represented by the planning system and transport system is illustrated. Any bottleneck to clog in any of these two systems or constraints resulting from the systems will adversely impede last mile delivery (Anderson et al., 2005; and Dablanc, 2007).

Built Environment Last mile delivery

Planning System

Figure 3.5: Inter-locking interactions between planning system, transport system and last mile delivery within the built environment

The attributes of the built and regulatory environment represented by the built environment (infrastructure and population) and planning and transport systems (regulatory environment) are as constraints described in Figure 3.2 and Table 3.5 and discussed individually thereafter. The constraints have can have either positive or negative impedance on last mile delivery and can exert either positive or negative impedance on last mile delivery. Positive or negative impact here is in terms of the correlation with availability or scale of the constraints. For example, the higher the population, the higher the impedance on last mile delivery.

Table 3.4: Built Environment, Planning System and Transport System Constraints

Constraints Impact Key studies on LMD Population + Burton (2000), Jenks et al. (2008), Eppell et al. density (2001), Chhetri et al. (2013) Proximity to + Ewing (1997), Jenks et al. (2008) activity centres Speed limit - Eppell et al. (2001), International Road Assessment Programme (2015) Toll +/- Eppell et al. (2001), He and Zhao (2014), Li, (2015) Traffic Lights + He and Zhao (2014), Luo et al. (2015), Tram lanes + National Transport Commission Regulations (2006) Planning Zones -/+ Victoria Planning Provisions Railway gates + National Transport Commission Regulations (2006) No of Lanes - Forkenbrock and Foster (1997), International Road; Assessment Programme (2015), Li and Choi (2002) Traffic count + Forkenbrock and Foster (1997), International Road; Assessment Programme (2015), Li and Choi (2002) Intersection + He and Zhao (2014), Luo et al. (2015) Bicycle lane + National Transport Commission Regulations (2006)

Last mile logistics is vital in achieving the key objective of the Victorian Planning Provision (2016). The provision states that “planning should ensure an integrated and sustainable transport system that provides access to social and economic opportunities, facilitates economic prosperity, contributes to environmental sustainability, coordinates reliable movements of people and goods, and is safe”. Ensuring that planning as an integrated and sustainable transport system provides access to social and economic opportunities, facilitates economic prosperity, contributes to environmental sustainability, coordinates reliable movements of people and goods, and is safe (VPP, 2019) is therefore the objective of land use and transport planning.

3.5.1 Population Density

Population density is a single most operational constraint that drives the efficient functioning of city systems (Dantzig and Saaty 1973; Breheny 1992; Beatley 1995; Jenks et al. 2008; Burton 2000). Population density is measured in terms of number of person per square kilometre. Table 3.5 and Figure 3.6 show population density across different local government areas in Melbourne. Population density within Metropolitan Melbourne tends to decrease with increasing distance from Melbourne city centre. However, there might be similar localised declining effect of population density from major business/commercial hub.

Population density of Port Philips, an inner city suburb, is about 4,803 people per square kilometre in 2016; whilst for an outer suburb of Yarra Ranges, it is 56 people per square km. Within the inner ring, Port Phillips has the highest population density, followed by Stonnington and Maribyrnong returning the lowest density (1,916 per square kilometre). Within the middle ring, Glen Eira has a population density of 3,005 per square kilometre, followed by Borrondara. Frankston has the lowest population density of 838 per square kilometre, which represents the lowest within the middle ring. In the outer ring, population density ranges from 428 per square kilometre for Casey and 56 per square kilometre in Yarra Ranges CC.

Table 3.5: Population density and distance of Councils from CBD

Population (2016) Population Density per Distance from Local Councils Square km CBD Melbourne 135,959 3675 2 Port Philip 100,863 4803 6 Stonnington 103,832 3994 13 Yarra 67,052 3353 4 Maribyrnong 59,406 1916 9 Banyule 113,696 1805 23 Bayside 83,504 2320 12 Boroondara 167,231 2787 19 Brimbank 194,319 1580 19 Greater Dandenong 123,965 954 28 Darebin 122,821 2317 14 Frankston 109,808 838 42 Glen Eira 117,199 3005 17 Hobson Bay 80,120 1233 22 Kingston 127,540 1402 20 Knox 141,408 1240 31 Manningham 107,079 948 21 Maroondah 96,132 1576 33 Monash 155,061 1891 23 Moonee Valley 105,615 2400 9 Moreland 130,531 2559 8 Whitehorse 139,549 2180 26 Casey 175,505 428 46 Cardinia 94,128 73 58 Hume 131,182 261 40 Melton 51,685 98 43 Mornington Peninsula 124,891 173 54 Nillumbik 57,932 133 25 Whittlesea 113,784 232 28 Wyndham 84,861 157 31 Yarra Ranges 137,113 56 37

Figure 3.6: Land area, population and population density of Local Councils in Metropolitan Melbourne.

3.5.2 Activity Centre

Activity centres are often termed Central Activity Centres, which represent a mixed use high intensity urban development area. They are often vibrant community hubs where people shop, work, meet, relax and often live.

The key objective of Activity Centres with the Victoria Planning Provision Clause 11.03-1S is to “encourage the concentration of major retail, residential, commercial, administrative, entertainment and cultural developments into activity centres that are highly accessible to the community. It seeks to have a focus for business, shopping, working, leisure and community facilities and provides different types of housing, including forms of higher density housing. Strategies within the area includes a reduction in the number of private motorised trips including delivery vehicles and generate high numbers of trips in highly accessible activity centres.

As a result of the creation of these highly concentrated urban forms, roads are getting narrower to create opportunities for public transport and cycling paths. These activity centres are imposing higher impedance to last mile logistics. The breakdown of the Activity centres associated with the development of the Melbourne as a compact city reveals 26 principal activity centres and 82 major activity centres (Appendix C). Five of these are located within

the inner ring with only three within the outer ring. The major activity centres are widely dispersed across the inner, middle and the outer rings.

In addition, population density within the Activity centres is much higher (12,034 person per square kilometre) within Melbourne CBD. This is due to the rapid rise in high-rise living. These areas have high dwelling density, intensive landuse, accessibility to multimodal transportation and pedestrian accessibility (Neuman 2005; and Chhetri et al. 2013). The implication of these characteristics potentially resulted in difficult access of freight vehicles and high impedance to last mile logistics.

3.5.3 Tollways

Toll as part of the built environment is a charge payable for the use of a road or bridge. This is a form of revenue collected from service-users for the construction and maintenance of roads or amenities. Within Metropolitan Melbourne, tolls are charged along sections of the Monash, the Bolte Bridge (City Link) and East link (Figure 3.7). These roads are bypass, which allow drivers to avoid delays because of congested roads, low-speed limit residential zone and higher traffic lights density. The roads with toll are within the inner and middle rings of Metropolitan Melbourne.

Figure 3.7: Toll Roads in Metropolitan Melbourne: (Source: VicRoads)

The toll roads are characterised by high speed of between 80km/h and 100km/h, and three- four lanes in one direction. Such roads are likely to make a delivery faster than going through areas with traffic lights. However, drivers are often frustrated with delays associated with the use of such toll road and the payment required to travel on the road. This can therefore be of

positive and negative impedance to last mile delivery, depending upon the way they impact route choices. In the calculation of last mile delivery impedance index, toll ways with high speed 3 – 5 lane travel in one direction, with road shoulder and service/exit lane are considered as positive impact on last mile delivery.

3.5.4 Speed limit

Speed limit is the maximum speed at which a vehicle may legally travel on a particular segment of a road (Road Management Act 2004; National Transport Commission Regulations 2006). It is often express as a sign along the road in number representing kilometre per hour in Australia.

The speed limits on the road network range from 39km/h to 110km/h. The areas with 40km/h and 50kph are local roads within areas designated as residential by the planning scheme and the Victoria Planning Provision and roads within the school zones. 2,364 (43.5%) segments of the roads are between 39km/h and 50km/h. Out of these roads segments, 156 (2.9%) of the total roads are 40km/h or less. The roads of between 60km/h and 70km/h represent 26% (1415) of the roads. Roads with between 80km/h and 110km/h represent 30.32% (1645) of the roads.

Speed as a road attribute affects the distance that can be covered within a specified time. Hence, high speed limits reduce the impedance to last mile delivery. Lower speed limits results increase travel time; while higher speed limits decrease in travel time (Li et al., 2006)

A total of 5,424 road segments are within Metropolitan Melbourne. The roads include declared roads and local roads. The declared roads are arterial roads identified as Road Zones 1 and 2 within the Victoria Planning Provisions. These arterial declared roads carried the major portion of last mile delivery vehicles.

3.5.5 Traffic lights

Traffic lights are road signals to direct vehicular traffic by means of coloured lights. Traffic lights are typically red for stop, green for go, and yellow for proceed with caution. In recent decades, increased use of intelligent traffic system has become more prevalent in many cities (Luo et al., 2015) particularly in densely populated part of the city (Transportation Research Board, 2000). The numbers of traffic lights within segments of roads are generally different.

More often traffic lights are more prevalent on roads of less than 80km/h. There are 4,556 traffic lights within the Melbourne Metropolitan roads. Over 80% of the traffic lights are located within the inner and middle ring. Uncoordinated and less synchronised traffic lights are often result in longer travel time for freight vehicles. The more the number of traffic signals, the higher the level of impedance to last mile delivery.

3.5.6 Landuse Zoning

Land use zoning is a planning tool for designating permitted uses of certain parcels of land -. Zoning within Metropolitan Melbourne are within the Victoria Planning Provision and included in each Planning Schemes of local councils. The basic planning zones designates land for agricultural, commercial, industrial and residential at varying degree

Section 30 of the Victoria Planning Provisions identified 6 major zones (Appendix D). These include Residential Zone, Industrial Zone, Commercial Zone, Rural Zone, Public Land Zone and Special Purpose Zone. The breakdown and the specific zones within the zones are provided in Table 6.3. Of particular interest within the zoning provision of the Victoria Planning Provisions are Residential Zone, Industrial Zone, Commercial Zone, and Special Purpose Zone within which Activity Centre is nested.

Different land use zones impose different level of impedance to last mile delivery. This is regulated through restriction related to the level of population density and the level of constraints resulting from the transport systems and controls. For example, the types of zone influence the hierarchy of Roads (Eppel et al 2001) which determines the speed limit. Specifically, residential zones, are characterised with relatively lower speed limit, fewer lanes, larger number of traffic lights and intersections/roundabouts, dedicated bus lanes, and higher to medium population density. Higher and harder restrictions in inner city areas such as parking, loading/uploading and multi-road use impact on last mile delivery. In contrast, areas outside residential zone and activity centres characterise different features such as higher speed limit, less traffic lights, availability of shoulder for parking of last mile delivery, in case of emergency.

Areas of higher speed limit especially with 80 km/h to 100 km/h will not be within the close proximity of school zones, will not have railway boom gates/rail crossing and tram routes.

These areas with higher speed limits will thus have lower impedance level to last mile delivery.

3.5.7 Traffic Count

Road congestion is a function of the road characteristics and the number of vehicles on the road at any particular time. This is measured from low to high level of congestion, reflected through the speed of vehicles’ movement per unit of time. High traffic congestion is characterised by slower speed, longer trip times and increased vehicular queuing. The determination of low or high congestion is determined by traffic count along the road segment. Traffic count per unit of time is therefore used as a surrogate to represent the levels of traffic/road congestion. For instance, a minimum of 10 vehicles along Acheron Way (Yarra Ranges CC) to a maximum of 96,000 along High Street section of the Monash Freeway (Monash CC) vehicles is accounted for within the Metropolitan Melbourne. The road segments recorded a total of 81,180,140 vehicles, which is an average of 14,967 vehicles per day. Trucks accounted for 5,809,827 with an average of 1071 vehicles per day representing 14% of the total vehicles. This is less than the 20 – 30% estimate of Christopher (2016). Generally, high vehicle volume occurs within the inner and middle rings with lower traffic volume within the outer rings Traffic count is largely driven by Traffic generation, which is a function of vehicle ownership and the usage of the vehicles for different journey from journey to work, social, educational purposes and freight deliveries.

The more the number of vehicles on the road, the higher the level of potential impedance on last mile. Road congestion is therefore a negative constraint that deters last mile delivery. .

3.5.8 Number of Lanes

Lane is one important element of the transportation network, which influences the traffic flow. A lane is a marked portion of the pavement for “vehicles travelling in the same direction as the driver but does not include a special purpose lane in which the driver is not permitted to drive” (The Australian Road Rules 2012, pp. 22). It is a part of the carriageway which is designed to be used by a single line of vehicles to control drivers and reduce traffic conflicts. Lanes could be one way in a single direction or two lanes in two directions or multi lanes in two directions. There are lanes that are designed only for buses and trams. The lane markings on multi-lane segments of the road distinguish each lane from the other.

The flow of traffic, speed and the capacity of the road is partly a function of the number of lanes on the road, which represents the road capacity. A two lane two direction road limit the speed and the number of vehicles that can traverse on the segment of the road. Consequently, the higher the number of lanes on a road segment, the lower the potential impedance of the road on last mile delivery. In contrast, higher the number of lanes, the availability of road shoulder and service lanes, the lower the level of delay to last mile delivery. For example, a road with 3 or more lanes will enhance better travel time and will be of lower impedance. The number of lanes in one direction within Metropolitan Melbourne ranges from 1 to 8 lanes (Figure 3.8).

Figure 3.8: Typical Melbourne road - Westgate Freeway (Source: Google maps)

3.5.9 Tram Lanes

Tram is a major public transport mode within the inner and middle rings of Metropolitan Melbourne. It comprises of approximately 250 kilometres of track which are spread over 24 routes. Tram tracks in Metropolitan Melbourne either form part of the road pavement or separate right of way (ROW) with tracks laid in mass concrete (Figure 3.9).

There are often conflicts and delay resulting from shared pavement where tram tracks forms part of the road surface (Road Management Act 2004). This is as a result of the priority given to trams over other vehicles especially at intersections and tram stops designated as safety tram zones. Separating of the ROW for trams clear of the roadway will limit this conflict and delay. Delay resulting from trams is seen as a constraint to freight movement; therefore the presence of tram on road is a negative impedance to last mile delivery.

Figure 3.9: Shared Tram and vehicle roadway (Source: Google maps)

3.5.10 Railway Boomgate or Level Crossing

Road and rail (train) are two other forms of land transportation. Both forms can be on the same grade or be separated. Level crossing is known as a key barrier to vehicular movement which cause significant road congestion. The word level crossing is used to describe the intersection of both rail and road on the same level in the Australian Road Rules (2012). At level crossing vehicles are required to stop and give way to train or tram approaching the crossing (Figure 3.10). Railway boomgates are used in many of such crossings. Boom gates and tram crossings are considered as negative constraint to last mile delivery.

In addition, roads in close proximity of railway crossing boomgate tends to have lower speed limits and the more likely to slow down the movement of freight vehicles. The removal of the level crossings being embarked by the Victorian Government will elevate this conflict and reduce the impedance resulting from the constraint. The removals of the boomgates are currently being undertaken. Figures 3.10 and 3.12 reveal the new sky rails at Clayton and Hughesdale. The removal of 50 of such level crossing boomgate will cost $8.3billion.

Figure 3.10: Railway boomgate in operation (Source: Victoria Level Crossing Removal)

Figure 3. 11: Clayton Road overhead rail pass (Source: Victoria Level Crossing Removal)

Figure 3.12: Poath Road, Hughesdale Rail overhead (Source: Victoria Level Crossing Removal)

3.5.11 Bicycle Lanes

Bicycle lanes are paths on the roadway specifically designated for cyclists. Where such bicycle path is considered from the initial conception of road construction, it does not constitute any delay to last mile delivery. However, many bicycle paths (Figure 3.13) results from cutting out from existing roadways. Such cutting reduces road width and number of lanes available for vehicular movement. Bicycle lanes as a constraint along the roadway is therefore a negative factor of last mile delivery impedance. Human behaviour associated with vehicle driving using the bicycle lanes are also a noticeable factor.

Figure 3. 13: A typical cut out bicycle lane (Source: Google maps)

3.5.12 Other Constraints

There are other variables such as road closure, low bridge height and human behaviour that affect the travel efficiency of last mile delivery vehicles. For example, there are implications on last mile delivery resulting from restrictions on local roads and residential roads by local authorities. This restriction are in the form of “No Trucks” restrictions on certain designated lanes or limits imposed by low height of bridges that prevent access to roads for last mile

delivery vehicles no matter how close or proximal such roads are to the pick up or delivery points.

3.6 Summary

In summary, this chapter has provided the background of the Theory of Constraints, the thinking process and Scenario Thinking that collectively laid the foundation of the theoretical framework to estimate last mile impedance and formulate LMD future scenarios. The chapter discusses the question what is, identifying the key constraints impeding last mile delivery. The discussion on the second principle of TOC with Scenario Thinking focuses on what to change, what to change to and how to cause the change. A five-stage Scenario Thinking method was adopted to understand the provision of last mile delivery and to build the possible LMD scenarios with ‘critical uncertainties’.

The chapter finally developed the theoretical framework that integrates TOC and ST to estimate current levels of LMD impedance and the required steps to formulate future LMD scenarios.

The research methodology including the approach adopted and data collection and analysis follows in chapter four.

CHAPTER FOUR

RESEARCH METHODOLOGY

4.1. Introduction

The preceding chapter established the theoretical framework that integrates Theory of Constraints and Scenario Thinking to formulate LMD scenarios. This chapter presents the research methodology to implement the framework by outlining the approach adopted, methods applied and the data collected. Discussion on Scenario Thinking as the key method in this study is presented to formulate the plausible scenarios to project the likely outcomes for the future of last mile logistics.

This chapter is organised into six sections. It begins with an all-encompassing discussion on the approach adopted in the research in section 4.2. Section 4.3 highlights the study area, while section 4.4 presents the research design that includes the integration of Geographical Information System in the methodological framework. The section also features the discussion on the selection of participants for the Scenario Thinking workshop and the research stages as well as the sources of the mappable data. Section 4.5 presents associated ethical consideration for the study. Section 4.6 provides the summary of the chapter.

4.2 The Approach

Approaches in social sciences research are based on their contrasting ontological, epistemological and methodological bases. These are in terms of related existence of a real and objective world; possibility of knowing this world and the knowledge forms; and the procedural mechanisms used to acquire that knowledge (Corbetta, 2003). This set of approaches, otherwise known as paradigm governs the key questions to be developed and the line of enquiry to be pursued (Filstead 1979; Guba 1990; Guba and Lincoln 1994; Mertens 2007). A paradigm is defined as a “set of linked assumptions about the world which is shared by a community of scientists investigating the world” (Deshpande 1983, pp.101). Paradigms as worldview, through ontology and epistemology helps researchers to identify what problems are worthy of exploration and what methods can be deployed to address these problems.

Ontology fundamentally questions ‘what is’ in relation to the nature of existence and reality (Crotty 1998) and emphasises the expectations that concern the very essence of the phenomena under investigation” (Burell and Morgan, 2005). Epistemology questions the theory of knowledge, how we know things (Porta and Keating, 2008) and the nature, sources and limits of knowledge (Klein, 2005). Methodology pinpoints the practices used to attain knowledge and defines the link between the methods and related paradigm (Healy and Perry 2000; Berth, 2002). This discussion on ontology and epistemology guides this research, the research design, and plan to be followed. Methods are a subset of methodology focusing on the techniques and procedures employed for data collection and analysis.

There are four associated archetype categories related to these three paradigms which Guba and Lincoln (1994) identified as positivism, realism, critical theory and constructivism. The summary of the four categories and the ontological, epistemological and methodological approaches are presented graphically in Figure 4.1 and in Table 4.1 (Guba and Lincoln, 1994 and Healy and Perry, 2000).

Methodology Paradigm

Grounded theory CONSTRUCTIVISM

In-depth interviewing and focus group REALISM (with an interviewer protocol) Theory- building Instrumental case research REALISM research: emphasis Survey and structural POSITIVISM on equation modelling meaning Survey and other POSITIVISM multivariate techniques

Theory-testing research: emphasis on measurement Figure 4.1: Range of methodology and related paradigms (Source: Healy and Perry, 2000)

Table 4.1: Paradigm and ontology, epistemology and methodology

Paradigms Positivism Constructivism Critical theory Realism Ontology Reality is real Multiple local Reality (virtual) Reality is ‘real’ and and specific shaped by but only apprehensible ‘constructed’ social, imperfectly and realities economic, probabilistically ethnic, political, apprehensible cultural and and so gender values, triangulation crystallised over from many time. sources is required to try to know it. Epistemology Findings are Created Created Findings true findings – findings – probably true - Researcher is Researcher is a Researcher is a Researcher is objective by participants’ participants’ value aware and viewing reality within what is within what is needs to through a ‘one- being being triangulate any way’ mirror investigated investigated perceptions collected. Common Mostly In-depth Action Mainly Methodologies concerns with a unstructured researcher and qualitative and testing of interviews, participants applied theory. Mainly participants observation quantitative quantitative observation, methods methods like action research survey, and grounded experiments and research theory verification of research hypothesis

(Source: Guba and Lincoln, 1994)

Positivism asserts that the only way knowledge can be authenticated is through positive verification. This is the position in studies of the pure sciences (Healy and Perry, 2000), where the scientist can totally disassociate from the object of study. Realism, critical theory and constructivism in contrast do not rely totally on quantitative approaches, but with relevant qualitative approach. Realities depend on the interaction between the interviewer (the researcher) and the respondent participants (Healy and Perry, 2000). Critical theory involves long-term ethnographic and historical studies of organisational processes and structures with subjective assumption. Knowledge in the critical theory is basically in social and historical routines and value dependent (Guba and Lincoln, 1994).

In social science research, the debate on the use of normative or realist, qualitative versus quantitative techniques with regard to answering research questions is common among researchers (Porta and Keating, 2008). Research methodologies are often associated with balancing the individual strengths and weaknesses of both the qualitative and quantitative techniques (McGrath, 1982:70).

Applying a pure quantitative positivist approach (logical positivism) or pure constructivism to this thesis will limit the intent of this research and will not be able to answer the research questions set for this thesis. Research findings resulting from application of positivist approach will be purely quantitative and thus lack insight into in-depth issues (Payne and Payne, 2004) of the social dimension of stakeholder involvement. Application of pure constructivism will also be qualitative and inadequate. Mapping and visualisation of the level of impedance on the transportation network will not be possible. The application of one of these two will be inadequate to map the potential last mile delivery impedance and construct future scenarios of last mile delivery in a changing urban setting.

The realist approach, which has the elements of both positivism and constructivism (Healy and Perry, 2000) is therefore appropriate to be adopted to carry out this study. Realism is concerned with investigation of social issues while still using a scientific approach. The use of Scenario Thinking (a qualitative approach) and large spatial data GIS (quantitative tool) for mapping and visualisation is consistent with the intent of this research. Combining both qualitative and quantitative methods makes this research to be innovative for applicability in last mile delivery studies, which are highly dependent on uncertain urban and regulatory environment.

4.3 Research Methodology

The research methodology adopted in this study provides the specific procedures, methods and analyses of data collected to estimate LMD impedance and build future LMD scenarios. The thesis utilised both primary and secondary data sources, including the data gathered through Scenario Thinking workshop. A study area context is first presented, followed by Scenario Thinking workshop and research design that explains the modus operandi of building LMD scenarios.

4.3.1 Study Area

Melbourne is the capital of Victoria and the second biggest city in Australia. Melbourne in this research administratively represents Melbourne CBD and the local government areas which spread from the Port Phillips Bay coastline into the hinterland of Dandenong ranges to the east, Macedon Ranges to the west, southward Mornington Peninsular and the Yarra Valley.

4.3.1.1 History and population growth

The city of Melbourne was founded in 1835 by a group of Tasmanian businessmen who formed the Port Phillip Association to find a settlement on Port Phillip Bay. The area was occupied by the Kulin nation which was made up of up to five Aboriginal language groups. By 1869, the population grew to 700,000 from the early 50 settlers due to the gold rush of 1851. It earns the position of a town in 1842 and was declared a city by Queen Vitoria in 1847. Figure 4.2 presents the map of the growth of Melbourne in 1868.

The population of Melbourne has continued to increase greatly. The population has increased from 3.33 million in 2001 to 4.85 million by 2016 (ABS, 2018). The city is projected to be above 5 million in 2019 overtaking Sydney as the most populous city in Australia. Adding more than 600,000 households by 2030 (ABS 2018). This increased population will continue to put pressure on the last mile logistics. The growth pattern of Metropolitan Melbourne is displayed in Figure 4.3.

Figure 4.2: Melbourne and Suburb (1868) (Source: https://www.hotpress.com.au/1852-map- of-Melbourne)

2 Population (million) Population 1

0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Figure 4.3: Melbourne Population growth 2001 to 2016

4.3.1.2 Location and local councils

The Melbourne region refers to as Greater Melbourne or Metropolitan Melbourne. Metropolitan Melbourne currently comprises 31 local government areas of approximately

8,836 square kilometres (Figure 4.4). These councils were made up of previous smaller cities and shire councils. These cities and shires are amalgamated in 1994. The 31 councils and previous cities amalgamated are presented in Table 4.2.

Figure 4.4: Metropolitan Melbourne within Victoria

Melbourne has experienced an extensive urban sprawl. Melbourne is therefore highly car dependant with over 3.6 million car ownership using over 22,320 kilometres of road. The city has an extensive road transport network of arterial and freeways that radiates east to west. These roads include the Monash Freeway and West Gate Freeway as well as Eastern Freeway. Toll operated roads in Melbourne are the Bolte Bridge, CityLink and Eastlink. Transport network within Melbourne appears radial and the major roads are managed by VicRoads; whilst the local roads are managed by local councils.

Table 4.2: Amalgamated cities to form the current 31 councils of Metropolitan Melbourne

Councils Amalgamated Cities Melbourne State of Victoria capital. Pot Philip Cities of St Kilda, South Melbourne and Port Melbourne. Stonnington Cities of Malvern and Prahran Yarra Cities of Richmond, Collingwood, Fitzroy and Northcote. Maribyrnong City of Footscray and parts of the City of Sunshine. Banyule City of Heidelberg and parts of the Shire of Diamond Valley and the Shire of Eltham. Bayside Cities of Brighton and Sandringham and parts of the cities of Mordialloc and Moorabbin. Boroondara Cities of Camberwell, Kew and Hawthorn. Brimbank Cities of Keilor and Sunshine. Greater Dandenong City of Springvale and City of Dandenong. Darebin Cities of Northcote and Preston. Frankston - Glen Eira City of Caulfield and parts of the city of Moorabbin. Hobson Bay City of Williamstown and the City of Altona. Kingston City of Chelsea and parts of the City of Mordialloc, City of Moorabbin, City of Oakleigh and City of Springvale. Knox - Manningham - Maroondah City of Ringwood and City of Croydon. Monash City of Oakleigh and the City of Waverley. Moonee Valley City of Essendon and parts of the City of Keilor. Moreland City of Brunswick, City of Coburg and part of the City of Broadmeadows. Whitehorse City of Box Hill and city of Nunawading. Casey City of Berwick with parts of the Shire of Cranbourne. Cardinia Shire of Pakenham, City of Cranbourne and parts of the Shire of Sherbrooke. Hume - Melton - Mornington Peninsula Shires of Flinders, Hastings and Mornington. Nillumbik - Whittlesea - Wyndham - Yarra Ranges Parts of the Shire of Lilydale, Shire of Healesville, Shire of Upper Yarra and parts of the Shire of Sherbrooke.

The 31 Local Government Areas within the Metropolitan Melbourne are classified into inner, middle and outer rings to evaluate the difference in last mile impedance. The classification is based on the distances of the Local Government Areas from the Melbourne CBD. Five local councils, consisting of Melbourne, Port Philip, Stonington, Yarra and Maribyrnong are within the inner ring. These are councils within 9 kilometres radius from the Melbourne City

Centre. Middle rings Councils consists of 17 local councils, located between 10 and 25 kilometre radius. The outer rings consist of 9 local councils that lie outside the 25 kilometre radius (Table 4.3 and Figure 4.5).

Table 4.3: Local Councils within Metropolitan Melbourne

Rings Local Councils Inner Ring Melbourne, Port Philip, Stonnington, Yarra and Maribyrnong. Middle Ring Banyule, Bayside, Boroondara, Brimbank, Greater Dandenong, Darebin, Frankston, Glen Eira, Hobson Bay, Kingston, Knox, Manningham, Maroondah, Monash, Moonee Valley, Moreland and Whitehorse, Outer Ring Casey, Cardinia, Hume, Melton, Mornington Peninsula, Nillumbik, Whittlesea, Wyndham, and Yarra Ranges.

Inner Ring Middle Ring Outer Ring

Figure 4.5: Metropolitan Melbourne in three rings

4.3.1.3 Urban Planning and Compact City Model

The containment of the Urban Growth Boundary and the Green Wedge policy towards the containment of urban sprawl resulted in a compact city model.

Following the partial success of the implementation of the compact city approach in Europe, Melbourne 2030 was introduced as a planning tool (Department of Infrastructure, 2002) to

curtain urban growth and caters for increased population growth. To implement a more compact city, Melbourne 2030 introduced the Activity Centre Policy and included it in the Victoria Planning Policy (VPP). The policy is to build up activity centres as a focus for high quality development, activity and living for the whole community; broaden the base of activity in centres currently dominated by retail and shopping to include a wider range of services over longer hours and restrict out-of-centre development; and locate a substantial proportion of new housing in, or close to activity centres and other strategic redevelopment sites that offer good access to services and transport (VPP, 2019).

A five broad activity centre classification (Figure 4.6) was introduced by the Melbourne 2030 policy. These include Central Activities District (Melbourne CBD), Principal Activity Centres, Major Activity Centres, Specialised Activity Centres, and Neighbourhood Activity Centre (Department of Planning and Community Development 2008).

Figure 4.6: Spatial Configuration of Activity Centres. (Source: Department of Planning and Community Development, 2008).

The breakdown of the Activity Centres associated with the development of Melbourne as a compact city reveals 26 principal activity centres and 82 major activity centres (Figure 4.7).

Figure 4.7: Location of Activity Centres within Metropolitan Melbourne (Source: Department of Infrastructure, 2011)

The classification of the activity centres defined the specific roles and functions that they play, preferred uses, the scale of development and link to public transport system. The emphasis of the policy centred on availability of public transport with little effort directed to freight movement within the compact city.

4.3.2 Key Case Studies

Case study approach is adopted in the research due to its ability to provide an in-depth and detailed study to the description and exploration of what happens, how and why (Willing 2001; Yin, 2004) in last mile delivery discussion. The approach is selected to provide a detailed understanding of last mile delivery at local level given the wide coverage of Metropolitan Melbourne.

Three case studies are selected to represent the local government areas across inner, middle and outer rings. These local governments are selected on the basis of their peculiar characteristics. The characteristics considered include distance from the CBD, population density, number of road segments, and location within each of the inner, middle and outer rings of Metropolitan Melbourne. These selected councils are Maribyrnong City (inner ring), City of Monash (middle ring) and Nillumbik Shire (outer ring). These councils also represent the east, north and south of the Metropolitan Melbourne. In addition, the selected councils perform different associated last mile delivery functions across Metropolitan Melbourne.

Figures 4.8 and Table 4.4 reveals the location of the local government within the Metropolitan Melbourne and a comparison of the land area, population density and number of road segments of the selected councils.

Nillumbik

Monash Maribyrnong

Figure 4.8: Location of Maribyrnong, Monash and Nillumbik within Metropolitan Melbourne.

Table 4.4: Key case study area comparison

Attributes Maribyrnong Monash Nillumbik Land area (km2) 31 82 435 Population density per square km 1916 1891 133 Road Segments 134 329 102 Distance to Melbourne CBD 9 23 25 Suburb Ring Inner Middle Outer

4.3.2.1 Maribyrnong City

Maribyrnong City Council is selected due to its proximity to Melbourne Port and it’s identification as a gateway to the western part of Melbourne Metropolitan areas. The City of Maribyrnong is located within the inner western suburb of Metropolitan Melbourne. The Maribyrnong River bounded the city to the north and east with Moonee Valley City located to the north east. The council shares boundary with Brimbank City to the west with Melbourne City located to the east. The West Coast Bay defined the southern boundary of the city, while the Hobson Bay City is located to the south. Generally, Maribyrnong City sits between Melbourne’s Docklands and port and the outer western industrial and residential areas and is currently experiencing rapid changes with significant new residential

developments now occurring. The city, with its land supply, major transport routes and accessibility to the port and airports, is a significant growth area in Metropolitan Melbourne. The changing land use patterns and the extent of new development within Maribyrnong over the past 10 years have significantly changed the visual appearance and urban form of the city. The dominance of the industrial character and image it held in the past has receded, and the city’s ‘renewal’ is bringing about a greater residential character and reputation.

The city’s proximity to Melbourne CBD allows for greater accessibility to employment, education, retail and business services. Several important east-west metropolitan road and rail transport routes traversed through the city with the freight transportation routes to the Port of Melbourne of state and national importance.

Close proximity of the city to the Port of Melbourne also results in significant adverse impacts on the local community due to congestion and heavy truck traffic on local roads. Activity centres such as Footscray Metropolitan Activities District and Yarraville are detrimentally impacted by heavy truck movements. Activity Centre Zone aimed at delivering of housing at higher densities towards the development of a compact Footscray Metropolitan Activity Centre was introduced via Clause 37.08 of the Maribyrnong Planning Scheme to the city.

4.3.2.2 City of Monash

The City of Monash is located to the south-eastern part of the Metropolitan Melbourne. The city is located approximately 23 kilometres from the south-east of Melbourne Central Business District. City of Whitehorse in located to the north while the City of Knox is located to the east. Immediately to the south are the City of Kingston and City of Greater Dandenong. To the west of the city are City of Stonnington, City of Glen Eira, and the City of Boroondara.

The City of Monash is a principally residential area, with large industrial, commercial and recreational spaces. The city has a total land area of 82 square kilometres. Monash University and the Monash Medical Centre are some of the best known landmarks within Metropolitan Melbourne. Monash Freeway transverse through the city. The city principally serves Wellington Road, North Road and Dandenong Road. The Glen Waverley and Dandenong

railway lines also passes through the city. Monash Freeway is an important arterial road that links the south east region with the other part of Metropolitan Melbourne.

4.3.2.3 Nillumbik Shire Council

Nillumbik is a shire council located on the northern urban-rural fringe of Melbourne, and located approximately 50 kilometres from the Melbourne Central District. The shire extends to the Kinglake Ranges with an area of 435 square kilometres. The shire generally bounded by the Yarra and Plenty Rivers and the Kinglake Ranges.

Manningham City and Banyule City bounded the shire to the south, while the City of Whittlesea is located to the west. Shire of Yarra Ranges is located to the east with Shire of Murrundindi located to the north.

Much of the shire is rural and is used for a combination of agriculture, rural living and conservation purposes. The population, however, is concentrated in the residential areas of Diamond Creek, Eltham, Greensborough, Hurstbridge, North Warrandyte, Plenty, Research and Wattle Glen.

4.3.3 Scenario Thinking Framework

This section presents the Scenario Thinking framework that includes the integration of the Scenario Thinking method, and GIS-based computation of last mile delivery impedance. Theoretically, the Scenario Thinking linked to the Theory of Constraints as discussed in chapter 3 drives the methodology adopted in this study.

Scenario Thinking is a participatory and democratic method that uses intuitive logic (thinking process). It is a group (stakeholder) thinking process. It facilitates sharing of knowledge and strategic conversation to create stories. It allows a deliberate participation where all viewpoints are considered equal and discussed openly.

The Scenario Thinking method is built on the first stage of Logistics Paradigm and the Thinking Process - that part of the first and the second principles of TOC. While the first stage of logistics paradigms allows the identification of constraints, thinking process allows viewpoints to be considered equal and all ideas expressed and discussed in an open and non- confrontational way. In Scenario Thinking, various steps are followed depending on whether

the approach is qualitative or quantitative. The qualitative approach is followed in this thesis. The five thematic Scenario Thinking stages (Figure 4.21) and associated tasks for the research are discussed in this section. The stages are abridged version of the 8 stage version proposed by Wright and Cairns (2011).

Theory of Constrraints Scenario Thinking

Stage 1 of Identification of last mile delivery constraints Stage 1 Logistics Paradigm

What to change? Clustering of attributes and defining of clusters Stage 2

Generating of Impact and Uncertainty matrix Stage 3 Thinking Process What to change to?

Scoping and Development of LMD Scenarios Stage 4

Stakeholder Power and Interest Matrix How to cause the Stage 5 change?

Figure 4.9: Seamless integration of Principles of Theory of Constraints and Scenario Thinking

4.3.4 Scenario Thinking Workshop

The Scenario Thinking workshop was facilitated by the researcher. Primary data are sourced through Scenario Thinking which is after the first and second principles of the Theory of Constraint (TOC) discussed in chapter 3. Scenario Thinking approach is used to gain an in- depth identification of built environment of transportation network and planning controls constraints by stakeholders in last-mile delivery. The Scenario Thinking workshop was held at RMIT.

Stakeholders in last mile can generally be categorised into three groups. These include the Administrators, Operators and the End Users (Russo and Comi 2011; Stathopoulos, et al. 2012). The use of Scenario Thinking in this instance is based on the ability to maximise the input of the experts involved in last mile delivery. The three categories informed the selection of workshop participants who participated in the Scenario Thinking workshop. These participants are involved in B2B last mile delivery.

The end-users are residents or consumers. These are individuals living, working and doing economic activities in the city. As a result of the vehicular movements including last mile

delivery trucks, they experience various levels of problems associated with delivery activities close to residential and retail areas. However, they also enjoy the benefit the delivery of their goods and will likely appreciate efficient and reliable delivery. The interest of this group (identified in this research as residents) is reduction or non-hindrance caused by the transportation of goods to congestion and other environmental issues.

City logistics operators identified by Russo and Comi (2011) as transportation operators deliver goods to customers. The aim of city logistics operators is to minimise their costs through efficient pickup and delivery. These city logistics operators are expected to provide high level of service at low costs. Their interest is minimum cost delivery to meet the customer needs and providing high quality delivery service with short lead time. This is in contrast to what the last mile delivery provider currently experienced.

City administrators endeavour to enhance the sustainable city development. This includes state and local governments. City administrators are interested in reduction of congestion and environmental pollution as well as the safety of the community. Therefore, they consider city transportation systems as a whole in other to resolve conflicts between the other stakeholders in last mile delivery. The interest of city administrators are to provide an attractive effective and efficient urban transportation to residents. Their key attention is towards minimisation of the negative effects of transportation and the maximisation of the net financial benefits.

City logistics service providers activities hinges on the interaction between stakeholders identified above. City administrators set complex restrictions through the planning process for the realisation of delivery trips. These restrictions includes certain permissible or prohibitive time slots on freight vehicles in pedestrianised areas. Also, they collect valuable mass data via traffic information systems. The interaction of the three stakeholders is presented as relates to last mile delivery in Figure 4.22.

Just in time delivery with little consequence

City Administrator Logistics Operator

Last mile Endusers delivery Attractive Urban area Reduction of congestion and environmental nuisance

Figure 4.10: Interaction and interest of stakeholders to cause the change in last mile delivery

4.3.4.1 Selection of Participants

Participants for the Scenario Thinking workshop were selected based on specific criteria and their understanding of the issue of discussion. The critical case sampling, a purposive sampling technique (Guarte and Barrios, 2006) that considers cases that are ‘particularly information rich’ in relationship to the questions under consideration, (Patton, 1987 and Yin, 2009) was adopted for the selection of participants. Purposive sampling technique utilised smaller number of participants (Ward et al., 2001) for enriched discussion on issues being studied. The advantage of smaller group is to give opportunity to participants to discuss and debate the focal issue (Greenbaum, 2011). Participants are drawn from local council, state government agency and last mile operators including retailers.

Fourteen participants were involved in the Scenario Thinking workshop. The Scenario Thinking workshop involves few open-ended and unstructured questions intended to elicit views and ideas from the participants. The view to be elicited from the participant in this instance is to validate the variables and issues in last mile delivery impedance. Four

participants represent the Administrator. These are made up of one participant from VicRoads (Victoria Government agency) and three participants that represent the Local Councils. The participants from Local Councils are made up of one Transport Planner, one each from inner ring (Maribyrnong), middle ring (Kingston) and outer ring (Wyndham). Six truck drivers and one Manager Logistics Operation represent the operators. Three participants represented the end-users. One representative each from market association from the Queen Vitoria Market (Melbourne), Springvale (Greater Dandenong) and Pakenham Community Market (Cardinia) representing the inner, middle and outer Councils.

4.3.4.2 Scenario Thinking - Setting of the Agenda

Setting the agenda entails defining the focal issue and the process. This stage starts with setting up the context within which the issues are defined and conceptualised. The agenda for this stage is founded on the existing literature and framed on key debates challenging the current last mile logistics practices. Understanding the constraints imposed by built and regulatory environment helps shape the scale and scope of this study. To allow active participation, every participant is provided in advance with printed Scenario Thinking workbook reference documentation that enables them to follow the full stage model.

Before participants arrived at the Scenario Thinking workshop, it is expected that they have some knowledge about the issues to be discussed. Participants at the Scenario Thinking have initial understanding of the issues relating to last mile delivery and have been involved in last mile delivery for the past 5 years in the capacity of Transport Planner, Driver, Logistic Manager or Retailer (Ewedairo et al., 2018). Before the workshop, the initial project brief, describing the purpose of the project and the activities to be carried out at the workshop were sent to participants (Wright and Cairns, 2011).

In setting the agenda, attempt was made to define and discuss “impedance” as applicable in this research. Also, the city logistics and last mile delivery operational definition and scope of the study were discussed including the negative perspective associated with last mile delivery. The discussion in the Scenario Thinking workshop has focused on the following key research questions.

 What are the constraints of the built environment that impedes last mile delivery within Metropolitan Melbourne?

 How will these constraint interact to impact last mile delivery in the next 10 years?  What are the possible and plausible future built environment scenarios given the interaction of the urban environment and last mile delivery within Metropolitan Melbourne?  How will the impact of these constraints be mitigated to limit its impact on last mile delivery?  Who will be in the position to effect this change?

Identification of last mile delivery Identification of the constraints of the Stage 1 constraints built environment that impedes last mile delivery within Metropolitan Melbourne.

Clustering and naming of How these constraints interact to impact Stage 2 Dimensions last mile delivery in the next 10 years.

Possible future built environment scenarios that are possible and plausible Generating of Impact and Stage 3 given the interaction of the urban Uncertainty matrix environment and last mile delivery within Metropolitan Melbourne.

How the impact of these constraints will Scoping and Development of LMD be mitigated to limit its impact on last Stage 4 Scenarios mile delivery.

Who will be in the position to effect this Stakeholder Power and Interest Stage 5 change? Matrix

Figure 4.11: Scenario Thinking steps and associated questions

4.4 Research Design

The research design follows a five stages process (Figure 4.9). Stage one identifies the last mile delivery constrains through literature review and Scenario Thinking and integrates the spatial data of the constraints to estimate last mile delivery impedance. The research seamlessly integrates the Theory of Constraints with Scenario Thinking. The first stage of

logistic paradigm and the thinking process informed the use of the Scenario Thinking approach.

Stage 1 brings the mapping and visualisation of LMD impedance using the key built and regulatory constraints which will inform the formulation of strategies to mitigate the impedance. Also included in stage 1 is the standardisation of constraints, development and mapping of LMD network impedance index.

The constraints identified through literature and Scenario Thinking workshop will answer research question 1: what are the key built and regulatory constraints that impede last mile delivery in an urban setting? Answering this research questions relates to what to change? This is one of the three questions in the thinking process.

Stages 2 and 3 focuses on clustering and defining of constraints, generating an impact and uncertainty matrix and identification of core dimensions that will shape the future. These clustered dimensions will answer research question 2: what are the core dimensions that drive the future of last mile delivery scenarios?

Two of the dimensions derived in stage 2 will be utilised to develop possible future scenarios in stage 4 to answer research question 3: what are the likely future outcomes of last mile delivery scenarios? Scenarios constructed during the workshop answered this question. Answering research question 3 relates to what to change to?

Stage 5 conducts the stakeholder analysis to ascertain power and interest of different stakeholders in last mile logistics. The results in stages 4 and 5 are combined with the mapping in stage 1 and assist in the development of last mile strategies to answer research question 4: what strategies can be formulated to mitigate the risk of last mile delivery failure? This research question relates to how to cause change?

Scenario Thinking Literature Review + Theory of Constraints Principles

Stage 1: Identification of last mile delivery constraints (RQ1)

TASK 1 Identification of built environment, planning and transportation systems constraints

TASK 2 GIS Standardisation and assessment of constraints for Overlay their potential last mile delivery impedance

TASK 3 Development and Mapping of the last-mile network impedance

Stage 2: Clustering of attributes and defining of clusters (RQ2)

Stage 3: Generating of Impact and Uncertainty matrix (RQ2)

Stage 4: Scoping and Development of LMD Scenarios (RQ3 and RQ4)

Stage 5: Stakeholder Power and Interest Matrix (RQ2)

Localised last mile delivery mitigation strategy (RQ4)

Figure 4.12: Research Process

4.4.1 Stage 1: Identification of Key Constraints

In this stage, participants were requested to explore and identify the constraints of built environment, planning and transportation network that impede last mile delivery within Metropolitan Melbourne. The process of impedance /identification was conducted first on individual basis, so that each participant can raise as many issues as they can think of within the allocated time. In doing this, participants are encouraged to think broadly with focus on the built environment, planning and transport related constraints and issues. All identified

constraints and issues are first recorded on an individual Post-it note. The constraints identified at this stage are logical and intelligible.

There are three tasks that need to be completed in this stage to generate an impedance index. These include identification of constraints, standardisation and estimation and mapping of last mile delivery impedance.

4.4.1.1 Task 1: Identification of mappable Last Mile Delivery Constraints

The identification of last mile delivery constraints were derived from literature review and Scenario Thinking workshop. The process of identification is discussed in the later part of the thesis. Task 1 links to the identification of constraints, which are extracted from the built and regulatory environment.

GIS method is integrated into stage 1 of the research method. The stage involves an identification of key last mile delivery constraints and standardisation, estimation for mapping last mile delivery impedance on road network.

Geographic Information System as a computerised data management system allows the combination of geographical information with other types of information thus allowing representation of various mappable constraints.

Spatial data from the VicData website are in lines, polygon and nodes format. For example, the transport network segments are lines format, while activity centres are polygon and traffic lights are nodes.

The buffer function is used as a zone around the transport network road segments in units of distance which other constraints are referenced to. A buffer distance of 0, 5 metres, 10 metres and 20 metres around the road network was used to join and relate nodes and polygon constraints.

Datasets and Data Processing

Spatial data downloaded are in ESRI shaped file format. In addition, metadata associated with each of the datasets are downloaded to provide clarity in the interpretation of the data. The currency of the data is 2016. The datasets are analysed using ArcGIS 10.5.1 software. The datasets collected include population density at SA1 level (ABS), speed limit, zones and

activity centres, tram, rail networks, traffic lights, number of lanes, railway boomgate, toll and intersections. Intersection is represented by the number of road segments (Table 4.6).

The spatial data representing the constraints are sourced from the from Australian Bureau of Statistics and Victoria Data website, Victoria's open data directory (https://www.data.vic.gov.au/data/group/spatial-data). The spatial data are collected from planning and transport sections of the VicData datasets directory.

Mapping of the last mile network impedance was carried out using ArcGIS software. Road Network Impedance levels are thematically mapped using natural break classification method that groups similar values by maximising the differences between classes. Generally, Thematic Mapping Technique is a process of showing geo-referenced data and variables. It uses cartographic symbols to represent spatial pattern with emphasis on a phenomenon under study.

Spatial data mapped to represent the constraints are presented in Table 4.5 and shown in Figures 4.12 to 4.19.

Table 4.5: Constraints and dimensions of data

S/No Constraints Data i. Population density Population Density ii. Zoning Zoning iii. Traffic counts Traffic Count iv. Activity Centres Activity Centre v. Intersection constraint Intersection vi. Speed limit Speed vii. Number of Lanes Number of lanes viii. Toll Toll ix. Railway Boom gate Railway Boom gate x. Traffic Lights. Traffic Lights. xi. Trams Tram routes xii. Proximity to freeway interchange Declared Road xiii. Distance from freight network Declared Road xiv. Location of adjacent arterial road Declared Road xv. Road Hierarchy Declared Road xvi. Bicycle Lanes Bicycle Lanes xvii. Proximity to other transport modes (Buffer to Railway, Railway, Tram and Road Tram)

Population density per square km High Density Medium Density Low Density

Figure 4.13: Population Density

Figure 4.14: Planning zones

Traffic Volume Low Medium High

Figure 4.15: Traffic Count within Metropolitan Melbourne.

Figure 4.16: Location of Activity Centres

Number of Lanes

1 – 2 3 – 5 6 - 9

Figure 4.17: Number of lanes within Metropolitan Melbourne

Figure 4.18: Speed limits

100

Figure 4.19: Location of Traffic lights within Metropolitan Melbourne

Figure 4.20: Rail lines including Tram Routes within Metropolitan Melbourne

101

GIS Data Analysis

GIS data analysis includes classification methods, and the applicable tools used for spatial data analysis

Classification involves the process of sorting or arranging of spatial data into groups. The processing is often done through representation through colour, size and symbol. The tools include representing features, categories, qualities, charts and multiple attributes with colours or symbols. The classification tools used to classify features is included in the tool. Methods of classification include manual, equal interval, defined intervals, quartile, geometrical interval and standard variation.

The natural break is used for classification in estimation of last mile delivery and mapping. The natural break classification tool is used to clarify attributes into classes based on natural groupings displayed in the data. This is carried out through identification of break points and selecting the best similar values maximising the difference between classes. The process determines appropriate boundaries where there are relatively big differences in data values. The classification is based on Jenks et al. (2008) Natural Break algorithm. The advantage of natural break is in its ability to be applied in data that are not evenly distributed, hence it allows changing to ordinal scale and labelling of data from small (low), medium or large (high).

While all the identified constraints were used for the other stages of the Scenario Thinking (clustering, impact and certainty and scenarios construction), only constraints with spatial data were used for task 2 of this stage which includes geo-processing and mapping of last mile delivery impedance.

The spatial datasets are geo-processed through the different tools built in ArcGIS to derive datasets for the research. The major functions used in this analysis are composite index, buffer, clip, spatial joins and relates and symbology.

A buffer is a useful proximity analysis tool. It is a method used to create a zone around map features, as an offset round the feature of interest. It is often measured in units of distance or time in a distance. Buffer can be created as a fixed valued distance around points, lines or polygons (Figure 4.10).

102

Figure 4.21: Typical buffer tool in ArcGIS (Source: ArcGIS)

The clipping function tool is used to create a polygon overlay. With the clipping tool, a polygon overlay is used on a target feature(s) to extract only the target feature(s) within the outlines area of the polygon overlay (Figure 4.11). The clipping tool is used to clip the road segments layer to the Metropolitan Melbourne administrative boundary.

Figure 4.22: Typical Clipping tool procedure (Source: ArcGIS)

Spatial joins functions have been used to match attributes based on spatial relationship of two features. The common locational attributes displayed by each of the attributes are the basic spatial relationship used for the matching. The FID of the spatial constraints data is used for spatial join in the processing of data in this research.

4.4.1.2 Task 2: Standardisation of Last Mile Delivery Constraints

Task 2 relates to mapping whereby these major constraints are evaluated, standardised and aggregated to generate last mile delivery impedance. Data are standardised using the maximum-minimum procedure. Data standardisation involves the process of converting a batch of data values into standardised units by removing the effects of the average size of the values in the batch and the size of the spread of values of data (Blind, 2004).

103

After the identification of constraints, spatial data representing these constraints are processed, transformed and standardised. This step involves the standardisation of available spatial data representing mappable constraints for assessment and computation of potential last mile impedance.

Data standardisation involves the process of converting a batch of data values into standardised units by removing the effects of the average size of the values in the batch and the size of the spread of values of data (Blind, 2004). The maximum-minimum procedure was used for standardisation of data collected in different measurement units. The maximum and minimum procedure is given as:

I = [(V-min)/(max-min)] x 100

where,

V is the observed indicator value (after imposition of bounds), and

I is the new, rescaled, index-number representation with a value ranging from 0 to 100.

The formula was reversed for data variables, which have a negative impact on last mile delivery. For these variables, the formula is:

I = [(max-V)/(max-min)] x 100

4.4.1.3 Task 3: Estimation and Mapping of the Last Mile Delivery Network Impedance Index

Task 3 proceeds to estimate and map last mile delivery impedance across the study area. These tasks are discussed below to provide the clarity on the methods and procedures applied to generate a metropolitan wide last mile delivery impedance index.

Estimation and mapping of the last-mile delivery network impedance index involves the generation of a new layer. This was done through an overlay function and application of composite index method. The use of composite index allows all the spatial layers to be added and divided by the total number of layers for the mapping of last mile delivery impedance index.

104

Composite index method captures complex realities of last mile delivery impedance that is difficult to be represented by an individual constraints. Designing and implementing a composite index in this case is reflective of the different constraints identified within the built and regulatory environment as related to last mile delivery. The formula for composite index is described as follows:

Last mile transport network impedance = V1 + V2 +V3 + . . . . Vn

Figure 4.20 reveals an example of GIS overlay functions of different layers of the spatial data for both the built and regulatory environment.

Figure 4.23: A Typical GIS overlay function (Source: The Department of Geography and GIS, University of Illinois)

4.4.2 Stage Two: Clustering and Defining of Dimensions

Stage 2 involves the approach of clustering the identified constraints, and defining the constraints into higher level dimension. The broader aim is to extract one set of connections to identify a smaller number of dimensions. Wright and Cairn (2011) recommended no more than 10-12 maximum of higher-dimension. Conversation was made about how each of the constraints relates to each other.

In this stage, participants brainstorm on the process of clustering constraints into small number of dimensions and describing the meaning and scope of the dimensions. This includes discussion on the positive and negative outcomes of the dimension, considering constraints included in the dimension.

105

Following the clustering of the constraints, the linkages and defining of the dimension was carried out so that every attribute is linked in some way to every other constraint via the dimensions, checking every relevant constraint. The dimensions are thereafter named after participants were satisfied with the linkages and interrelationships among constraints. Each dimension is recorded on a Post-it® note.

4.4.3 Stage 3: Impact and Level of Uncertainty

This stage proceeds with placing the clustered dimensions on Post-it note along the full length of the impact axis, ranging from low impact to severe impact in the future. While the certainty axis has same scale ranging from low uncertainty to high uncertainty. Figure 2.23 illustrates the two dimensional configuration of these two scales. Participants were then requested to positioning the extracted dimensions relative to each other through a process of discussion and consensus on the perceived degree of impact of each on the key issue. These dimensions were then relocated relative to the assigned positions along the two-dimensional space created by high/low certainty and high/low impact axes. The placing was done through the process of negotiation.

Generally, impact is level of control and the effect that something has on a situation or person, while certainty is a state of being confident without any doubt about something. In this instance, impact is considered in terms of the effect of stakeholders on last-mile delivery performance, while certainty is the levels of confidence about the likely impact. Certainty here is not considering whether or not the outcomes of the dimensions will occur, but how certain about what will be the impact of imbedded constraints in the dimensions. Where there is wide discrepancy in the views on the impact of a constraint, such constraint will be of high uncertainty.

Two factors A and B, independent of each other will form the two dimensions for the construction of scenarios. Factors A and B will be the two dimensions with the highest impact with the greatest uncertainty. It is around the uncertainties of these two dimensions that different scenarios are constructed (Figures 4.24 and 4.25).

106

Low Uncertainty

Low Impact High Impact

High uncertainty

Figure 4.24: Impact and uncertainty matrix on horizontal scale (Source: Wright & Cairns, 2011)

Low Uncertainty

Low Impact High Impact

A B

High Uncertainty

Figure 4.25: Vertical movement of dimensions within the Impact and uncertainty matrix (Source: Wright & Cairns, 2011)

107

4.4.4 Stage 4: Scoping the Future Last Mile Delivery Scenarios

Drawing upon the previously determined polar outcomes of Dimension A and B (from Stage 3), the group proceeded to consider how the sets of conditions (bestA1/worstA2 and bestB1/worstB2 interact with each other in four different combinations (A1/B1, A1/B2, over time; as a set of tools for making sense of complexity and ambiguity and A2/B1 and A2/B2) to produce what, in simplistic terms, might be described as best/best, best/worst, worst/best and worst/worst world conditions (Figure 4.26).

At this stage, the group brainstorms the descriptors of such future ‘world’ of last mile delivery considering all of the polar outcomes for all constraints in Stage 3 and considers how best they fit within the scenario outlines. At this stage, discussion takes place about the inter- connection between constraints and it is seen that there are cause/effect relationships between factors which will make some linkages immediately credible and others nonsensical. These combination descriptors form the basis of the four scenarios developed at the ST workshop. The scenarios frame the expected possibility of the future and the links through different interest of stakeholders.

The scenario outlines consist of a set of descriptors of four possible futures (Wright & Cairns, 2011) derived from interaction of the two dimensions A and B through building of logical and consistent storylines to reflect the current to the future of last mile delivery.

108

Extremes Extremes A1 +B2 A1 +B1

Best/Worst Best/Best

A FACTOR FACTOR

Extremes Extremes A2 +B2 A2 +B1 Worst/Wors Worst/Best t

FACTOR B Figure 4.26: Four quadrants of scenario building (Source: Wright & Cairns, 2011)

Scenario scoping follows sets four phases combination of the high level dimensions to present 4 scenario storylines. The interaction of the dimensions at their best and worst combinations was carried out. The term A1/ B1 are at best conditions, while A2 /B2 are at the worst conditions. The ideal world conditions will be when the constraints clustered in the dimensions are all in their best: that is clustered in A1/B1 which last mile delivery will aspire. However, this will change with A1/B2 or A2/B1. The worst expected outcome will be A2/B2, when last mile delivery will be at its worst.

4.4.5 Stage 5. Stakeholder Analysis

This stage considered the range of stakeholders who can affect the last mile delivery to achieve the ideal situation highlighted in the previous stage. The extent of power and interest of such stakeholders in last mile delivery were considered. The power and influence of the stakeholders are to effect change in reducing last mile delivery impedance. What is of note is that power and interest may change.

109

Stakeholders’ analysis represents the process of answering the third question in thinking process. The question: how to cause the change? and who are the people to could potentially trigger the desirable change? There are numerous actors who are directly or indirectly involved in last mile logistics (Wolpert and Reuter, 2012). They are in a position of initiating the change in last mile logistics with the intent to generate positive outcomes. These actors are identified as stakeholders in last mile delivery discussion, (Rao et. al. 2015; Bozzo et al. 2014; Taniguchi 2014; Crainic et al. 2009; Stathopoulos 2012; Aized and Srai 2014) who are likely to shape the future of last mile logistics in an urban setting.

The stakeholder was matrix divided into ‘X’ and ‘Y’ axis with four quadrants. Each quadrant was created to represent various categories of stakeholders. Wright and Cairns, (2011) labelled them as Context Setters, Players Crowds and Subject Power and interest of stakeholders ranges from low to high (Figure 4.27).

High Power High Power High Interest Contest Players Setters Involved actors Unaffected

Crowd Subjects Unaffected bystander Involved bystanders

Low Power High Interest Low Interest Figure 4.27: Stakeholder power and interest matrix (Source: Wright & Cairns, 2011)

The Context Setters represent stakeholders with high interest, but low interest. However, they can influence the overall future of last mile delivery and they are not affected directly by policies and decisions resulting from their actions (Ackermann and Eden 2011) described as the unaffected by Wright and Cairns (2011).

110

The Players are the involved actors (Wright and Cairns 2011). They have high power, but often with low interest. This is given that their activities are not mainly directed at last-mile delivery, but decision from their actions impact last mile delivery.

The Crowds represent the unaffected bystanders and they have low power and low interest in last mile delivery operations. The interest of the crowd is only in the delivery of goods which they ordered (Ackermann and Eden, 2011). Their interest and power can be raised to influence policies for timey last-mile delivery. Considerable delay resulting from last-mile delay may have consequences and change the interest they have in last-mile delivery policies.

Subjects are stakeholders with high interest, but with low power. They are identified as involved bystanders (Wright and Cairns 2011) affected by decisions on last mile delivery and are significant stakeholders (Ackermann and Eden 2011).Last mile policies and regulation affect their operation as a result cost of delivery or delay in delivery.

At this stage, discussion centred on defining who stakeholder in each of the matrix and what concerns them, their interest and power including how the power are exercised. This stage also includes opportunity for empathy, allowing participants to put themselves in the position of other stakeholder.

Also at this stage, strategies to mitigate the potential risk of delays or increased delivery costs proposed to tackle the potential challenges associated with last mile delivery. Following from the stakeholder analysis, the strategies proposed will take into consideration how each the administrator and end users can assist a more efficient and effective last mile delivery.

4.5 Ethical Considerations

Ethical consideration is an important part of any research. It aims at ensuring that the interests of participants in the research are not compromised or taken for granted. Hence participants should not be hurt in any form, either through breach of anonymity or confidentiality (Gray, 2013 and Bryman and Bell, 2015). The principles of ethics such as honesty, integrity, and respect for others, which are universally understood and generally accepted (Bryman and Bell, 2015) is applicable to this research.

Researchers have seven legal and ethical responsibilities to meet when conducting data collection with voluntary participation without any coercion. Participants were assured of

111

informed consent, no harm, confidentiality, anonymity and privacy, respect and abidance by the conditions as contained in the signed agreement (Bogdan and Biklen 2007; Bryman and Bell 2015). Voluntary participation implies that participants can exercise their rights to be involved in the survey and that participation was optional.

To conduct the Scenario Thinking, the voluntary requirements of the participants were emphasised. Would-be participants had the opportunity to examine the issues to be discussed during the Scenario Thinking before making their decision to become involved in the research. Participants were assured of the voluntary and anonymous aspects participation and that could withdraw themselves and any unprocessed data concerning them at any time, without prejudice. Participants involved in this research were informed that they were able to withdraw partially or completely at any time.

Administratively, the participants had to provide signed consent which also related to the use of information provided. Regarding the ethics associated with data collection through Scenario Thinking, participants signed the documents to indicate implied consent. Implied consent means that when the would-be participants would contribute during the Scenario Thinking discussion (A. P. Association, 2002).

This research does not discuss any personal sensitive issues and no name of participant is recorded, hence the confidentiality of the participants is maintained. The privacy of participants, the confidentiality of data provided by them, and their anonymity are protected and maintained.

On the researcher’s part, efforts are made to ensure that objectivity in data analysis is maintained, and also to ensure that the data collected is not misrepresented. All information collected is strictly confidential and can be accessed only by the researcher and his supervisors. Participants are assured that there is no perceived risk outside their normal day- to-day activities. All data will be kept securely at RMIT University for a period of five years before being destroyed.

This research, therefore, has fulfilled all ethical considerations governing the conduct and operation of this research. Therefore, it has been considered as posing a Negligible Risk to participants. The Risk Assessment Checklist was approved and accepted by the Business CHEAN RMIT University panel on 10 March 2017.

112

4.6 Summary

This chapter developed a methodological framework to conduct this study. This chapter presented the prevailing paradigms and evaluated their merits and demerits in developing effective research approach and design. The realist paradigm allowed the integration of GIS (quantitative) with Scenario Thinking (qualitative) for understanding the potential last mile delivery impedance and future last mile delivery scenarios. The case study of Metropolitan Melbourne as a growing compact city model provided the spatial foundation for explaining the likely impacts of the built and regulatory environment constraints on last mile delivery. In addition, the sources of primary and secondary data, GIS methods of analysis and the scenario workshop procedures including how participants were selected are discussed in the chapter. Also, the chapter presented the five-stage approach of the research design.

The analysis and discussion of the stages of the thesis are presented in chapters 5 and 6.

113

CHAPTER FIVE

IDENTIFICATION AND ESTIMATION OF LAST MILE DELIVERY IMPEDANCE

5.1 Introduction

This chapter aims to identify the key constraints of built and regulatory environment to estimate and map last mile delivery impedance in Metropolitan Melbourne using GIS. It focuses on stage 1 of Scenario Thinking method that enabled the identification of both spatial and non-spatial constraints through the consultation process via the ST workshop. Finally, it presents the results of the statistical analyses conducted to test whether LMD impedance varies across different attributes of urban form and planning controls.

The chapter is organised into five sections. Section 5.2 discusses stage 1 of the research design, specifically, the three tasks involved in the mapping of LMD impedance levels. Section 5.3 discusses result of last mile delivery impedance of the overall Metropolitan Melbourne, while section 5.4 proceeds with small area analysis using three case studies, each representing the inner, middle and outer rings of Metropolitan Melbourne. The section also includes an overall discussion at a localised level for Maribyrnong, Monash and Nillumbik Councils, while the summary of the chapter is presented in section 5.5.

5.2 Stage One: Identification and Estimation of Last Mile Delivery Impedance

This stage involves three tasks to be carried out as a part of the research design as shown in Figure 5.1. The participants at the Scenario Thinking workshop through intuitive reasoning identified built environment, planning and transport systems constraints to last mile delivery. These constraints are the key drivers with the potential to change the future outcomes of last mile delivery. The tasks include: the identification of built environment, planning and transport network systems constraints; standardisation and assessment of constraints to convert values on a standard range and for their potential impact on last mile delivery

114

impedance respectively; and estimation and mapping of last mile delivery network impedance.

Scenario Thinking Literature Review + Theory of Constraints Principles

Stage 1: Identification of last mile delivery constraints (RQ1)(RQ1)

TASK 1 Identification of built environment, planning and transportation systems constraints

GIS TASK 2 Standardisation and assessment of constraints for Overlay their potential last mile delivery impedance

TASK 3 Development and Mapping of the last-mile network impedance

Figure 5.1: Stage 1 of the research design

5.2.1 Task 1: Identification of Last Mile Delivery Constraints

Stage one of Scenario Thinking workshop enabled the identification of key constraints. Participants have identified a total of 34 different constraints representing the built environment, planning and transport network systems. While some of the constraints are mappable, others are non-mappable because of their behavioural constraints. In addition, there are some constraints, though mappable, but no spatial data is available to include them in the mapping. Out of the 34 constraints, only 17 are mappable with spatial data available within the Australian Bureau of Statistics and the Victoria data website. Table 5.1 presents the breakdown of the identified constraints with an indication of mappable attribute with geographic coordinates.

The identified constraints reflect the diversity of views of participants and reveal the important constraints to each category of participants. While there is consensus with some of the constraints listed, participants hold divergent views on some of the constraints identified through the Scenario Thinking workshop. Importantly, four constraints identified by the

115

drivers are not considered significant by the administrator and the retailer in the Scenario Thinking workshop. These constraints include: intersection constraints, roadwork restrictions when road works have been completed, lack of drivers’ knowledge on how to merge and road design afterthought. Specifically, drivers are frustrated when traffic work restrictions are still on the road after the completion of construction works. Generally, the Scenario Thinking identified toll as a major impedance. While this is not applicable to large operators, small scale operators identified it as major constrain and drivers tends to avoid toll as much a possible given the high cost of toll.

On the other hand, the participants representing the administrators identified the following as the constraints of impedance to last mile delivery apart from the ones identified by the drivers. Surrounding areas and vehicle ownership rates; population density; the year of land subdivision; change in road hierarchy and change in number of lanes; proximity to freeway interchange; and distance from freight network.

The constraints and the spatial data for the estimation and mapping are listed in Table 5.1. The constraints are: traffic count, population density, zoning, activity centre, intersection, speed, number of lanes, declared road, presence of railway boomgate, traffic lights, bicycle lane and tram routes. These constraints are derived from the Scenario Thinking workshop with validation through literature.

116

Table 5. 1: Key constraints identified by the workshop stakeholders

Constraints Mappable Spatial Data 1 Loading and Unloading area. Yes N/A 2 Number of available loading and unloading area. Yes N/A 3 Parking restrictions Yes N/A 4 Surrounding area and vehicle ownership No N/A 5 Traffic counts Yes Traffic Count 6 Population density Yes Population Density 7 Year of land subdivision Yes N/A 8 Zoning Yes Zoning 9 Size of shopping Centres Yes N/A 10 Number of shops (traffic generation and parking No N/A spaces) 11 Activity Centres Yes Activity Centre 12 Intersection constraint Yes Intersection 13 Speed limit Yes Speed 14 Road Network Capacity No N/A 15 Change in road hierarchy and change in no. of No N/A lanes 16 Road Closure. No N/A 17 Speed limit and change in speed limit No N/A 18 Road width No N/A 19 Number of Lanes Yes Number of lanes 20 Abrupt change in number of lanes No N/A 21 Proximity to Freeway interchange Yes Declared Road 22 Distance from freight network Yes Declared Road 23 Location of adjacent arterial road Yes Declared Road 24 Road network design and alignment (road No N/A geometry) 25 Road Hierarchy Yes Declared Road 26 Presence of Toll Yes Toll 27 Presence of Railway Boom gate Yes Railway Boom gate 28 Traffic Lights. Yes Traffic Lights. 29 Road design afterthought No N/A 30 Bicycle Lanes Yes Bicycle lane 31 Trams Yes Tram routes 32 Proximity to other transport modes (Railway, Yes Railway, Tram and Tram) Road 33 Lack of drivers understating of road merge No N/A 34 Road wok restrictions when there is no road works No N/A

5.2.2 Task 2: Assessment and Standardisation of Spatial Constraint Data

The second task is the standardisation of the spatial constraint variables, which were collected in different metrics. A range of GIS functions were used to convert raw data to comparable metrics or to derive new set of data variables (e.g. buffer around school zones). The spatial

117

join toolset of ArcGIS was used to string together the datasets. A less than 5 metres, 5 to 10 metres, 10 to 20 metres and greater than 20 metres buffers were extracted using a buffer function. Five metres join distance was also applied to traffic lights to avoid double counting, while 20 metres spatial join distance was used for activity centres. A farther buffer join distance can result in overlapping and double counting of traffic light, however, 5 metres may not capture adjoining activity centre. Also tram routes and location of railway boom gates requires accurate intercepts, hence a zero buffer distance was utilised for that constraint join.

Table 5.2: Built environment, planning system and transport system constraints

Constraints Impact Type of Variable Measurement Unit on LMD Population Density + Ratio No. per square kilometre Proximity to + Interval Distance measured in Activity Centres kilometre Speed limit - Interval km per hour Toll +/- Nominal Yes or No Traffic Light + Nominal Number Tram Lanes + Nominal Yes or No Planning Zones -/+ Nominal Locational Railway boom gate + Nominal Yes or No Number of lanes - Nominal Number Traffic Count + Nominal Number Intersection + Nominal Number Bicycle lanes + Nominal Yes or No

Following from above, clip, buffer and spatial join functions were used to associate all the datasets together to form one single data base. All other datasets are related and join to Declared Road dataset. The datasets are merged using the Target_FID of each attribute. Each data variable is imported and integrated into the ArcGIS using the maximum-minimum procedure.

The identified constraints are examined in relation to road segments as the independent variables. A segment is a stretch of road between two intersections. Each segment can thereafter be separated into sub segments as a result of intersection, speed limits and traffic lights. The number of separated segments is considered as a ratio of the overall length of the

118

road segments which form different levels of impedance to last mile delivery. For example, the higher the number of traffic lights and intersections are expressed in a ratio of the overall road segments.

The new set of values were then derived by subtracting the minimum value in the distribution from each observed value for each data layer (variable) and expressed as a percentage of the difference between the maximum (max) and minimum (min) values in the distribution

The measures are then calculated in such a way that the higher the value of the component variables, the higher the levels of last mile delivery impedance (and vice versa).

The type of variables including the impact of different sets of constraints on transportation network LMD listed in Table 5.2. It is important to understand the negative or positive orientation of variables for their impact on last mile delivery, type of variable and measurement unit. Table 5.2 presents the information on measurement scale and the positive or negative impact of the constraints on last mile delivery.

5.2.3 Task 3: Estimation and Mapping of the Last Mile Delivery Impedance Index

Once the data variables are standardised and converted into same unit of 0 - 100, an overlay function was employed to generate a new layer of last mile delivery impedance. A simple additive method is used to compute a composite index whereby each of these layers are added and divided by the total number of layers.

Mapping of the last mile network impedance was carried out using ArcGIS through an overlay function is GIS. Road network impedance levels are thematically mapped using natural break classification method that group similar values attached to road segments by maximising the differences between classes. Impedance scores were categorised into three groups of high, medium and low using thematic mapping method. Visualisation enabled spatial variability of last mile delivery impedance to be mapped and the key transport network hotspots of high LMD to be identified.

A total of 5,424 road segments are within Metropolitan Melbourne. These are declared roads, which are arterial roads identified as Road Zones 1 and 2 within the Victoria Planning Provisions. These arterial roads carried a significant portion of B2B last mile delivery vehicles. The roads segments are spread across Metropolitan Melbourne.

119

The calculated impedance index ranges from 0 – 1, where 0 represents no impedance at all and 1 represents highest level of impedance to last mile delivery. Using the natural break classification, impedance level are grouped into five major categories, namely 0 – 0.450 (very low), 0.451 – 0.512 (low), 0.513 – 0.576 (medium), 0.577 – 0.800 (high) and 0.801 – 1 (very high). Table 5.3 presents the breakdown of the impedance of road segments. In total, 28.1% (522 segments) represents very high, 16.8% (911 segments) high, 25.7% (1394 segments) as medium, 11.8% (642 segments) as low and 17.6% (956 segments) as very low.

The impedance levels are presented in Table 5.3 and spatially represented in Figure 5.2

Table 5. 3: Road segment with level of impedance Impedance Level % of Road Segments Very High 28.1% High 16.8% Medium 25.7% Low 11.8% Very Low 17.6% Total 100%

28.1% representing 1522 road segments are categorised as very high impedance level. These road segments cut across all Metropolitan Melbourne. 48%, 27.8% and 7.6% are located within the inner, middle and outer rings respectively representing 880 road segments, 504 road segments and 138 road segments. Road segments within the inner rings accounts for over one third of very high impedance road segments.

120

17% 28% Very High High 12% Medium Low 17% Very Low 26%

Figure 5.2: Last mile delivery impedance index across road segments

The levels of impedance are collapsed to council level to calculate mean impedance level of the road segments. The mean impedance level are categorised from very high to very low with impedance level from 1 to 0. 1 represents very high road segment, while 0 represents very low impedance. Calculated impedance on the road segments ranges from 0.954 (very high) to 0.401 (very low) (Table 5.4).

Table 5.3 reveals the mean impedance index for each council and the rank of each council from very high to very low ranked from 1 to 31, where 1 represents highest impedance and 31 represents lowest. Values in column 4 in Table 5.3 present the number of segments within each council. Columns 5 and 6 present the impedance level and rank of the councils respectively.

Generally, councils within inner rings have higher levels of impedance; while most councils within the outer ring are of low impedance. This could be simply attributed to geographic positioning relative to key commercial activity hubs and the characteristics of suburbia. Impedance levels therefore tend to increase with distance from the councils to the CBD. Melbourne City within the inner ring has the very high level of impedance, while Nillumbik Shire has the lowest level of impedance. All the Councils within the middle rings and 4 councils within the outer rings are of medium impedance while 5 councils within the outer ring return low impedance level.

121

: Calculated : within Melbourne. level Metropolitan impedance of 3

5. Figure

122

Table 5.4: Number of road network segments and impedance rank

Councils Population Distance Location Road Mean Impedance Levels of Density per from Segments Impedance Rank LMD square km CBD Impedance Melbourne 3675 2 Inner Ring 221 0.954 1 Very high Pot Philip 4803 6 Inner Ring 76 0.818 5 Very high Stonnington 3994 13 Inner Ring 54 0.819 4 Very high Yarra 3353 4 Inner Ring 82 0.891 2 Very high Maribyrnong 1916 9 Inner Ring 134 0.864 3 Very high Banyule 1805 23 Middle Ring 72 0.475 25 Low Bayside 2320 12 Middle Ring 43 0.488 24 Low Boroondara 2787 19 Middle Ring 61 0.538 18 Medium Brimbank 1580 19 Middle Ring 347 0.597 7 High Greater Middle Ring Medium Dandenong 954 28 282 0.524 21 Darebin 2317 14 Middle Ring 75 0.531 20 Medium Frankston 838 42 Middle Ring 222 0.578 11 High Glen Eira 3005 17 Middle Ring 92 0.576 12 Medium Hobson Bay 1233 22 Middle Ring 122 0.753 6 High Kingston 1402 20 Middle Ring 299 0.515 22 Medium Knox 1240 31 Middle Ring 316 0.539 17 Medium Manningham 948 21 Middle Ring 155 0.512 23 Low Maroondah 1576 33 Middle Ring 178 0.596 8 High Monash 1891 23 Middle Ring 343 0.555 14 Medium Moonee Middle Ring High Valley 2400 12 84 0.594 9 Moreland 2559 14 Middle Ring 51 0.471 26 Low Whitehorse 2180 26 Middle Ring 207 0.540 16 Medium Casey 428 46 Outer Ring 249 0.586 10 High Cardinia 73 58 Outer Ring 221 0.550 15 Medium Hume 261 40 Outer Ring 396 0.574 13 Medium Melton 98 43 Outer Ring 164 0.430 28 Very low Mornington Outer Ring Very low Peninsula 173 54 213 0.426 29 Nillumbik 133 25 Outer Ring 100 0.401 31 Very low Whittlesea 232 28 Outer Ring 154 0.45 27 Very low Wyndham 157 31 Outer Ring 184 0.538 19 Medium Yarra Ranges 56 37 Outer Ring 227 0.411 30 Very low

123

1.2

0.8

0.6 dancelevel

0.4 Impe

0.2

Knox

Yarra

Casey Hume

Melton

Bayside Darebin Monash

Banyule

Kingston

Cardinia

Glen Eira

Moreland

Brimbank

Nillumbik

Frankston

PortPhilip

Melbourne

Wyndham

Whittlesea

Maroondah Whitehorse

Boroondara

Stonnington

Hobson Bay

Maribyrnong

Manningham

Yarra Ranges

Moonee Valley

Greater Dandenong Mornington Peninsula Very High Low Medium Very Low High Figure 5.4: Mean last mile delivery impedance within Metropolitan Melbourne

Correlation test reveals a strong positive relationship between mean impedance level and population density, with a correlation coefficient of 0.662. However, there are strong negative relationship between population density and distance from CBD, and road segments. The analysis reveals a very strong negative (-0.809) between population density and distance from the Metropolitan CBD. The results are statistically significant at 0.01.

While there are positive relationship between distance from CBD and road segments (0.473) there is a negative correlation between mean impedance level and distance from Metropolitan CBD (-564). Both correlation coefficient are statistically significant at 0.01.

Furthermore, negative correlation exists between mean impedance and road segments. A weak correction coefficient of -0.139 is recorded between road segments and impedance level across Metropolitan Melbourne.

Figures 5.5, and 5.6 graphically present the relationships between mean impedance, population density, road segments and distance from Metropolitan CBD.

124

Figure 5.5: Population density and impedance level relationship scattered diagram

Figure 5.6: Impedance level and distance from Metropolitan Melbourne CBD relationship scattered diagram

125

5.3 Metropolitan Last Mile Delivery Wide Analysis

In this study, the analysis of last mile delivery impedance is conducted at two different spatial scales: metropolitan wide analysis and local area level analysis. This is because the planning implications and strategies to mitigate last mile delivery impedance would vary at different spatial scales.

The analysis reveals a significant difference in transportation network impedance levels across Metropolitan Melbourne. This spatial variability indicates the differences in network capacity and planning controls imposed through land use regulation and population density variability. The potential last-mile impedance on transport network is estimated and mapped which reveals that different segments of the transport networks have different impedance scores, which are attributed to hindrances imposed by the constraints of the transport network on freight movement.

18.3 % representing 911 road segments are categorised as high impedance level across Metropolitan Melbourne. 18.3%, 19.38% and 12.64% are located within the inner, middle and outer rings respectively. These represents 331 road segments, segments and 229 road segments within the inner, middle and outer rings respectively.

Roads with medium level of impedance are 23.7% (430 road segments) within the inner ring, 30.5% (551 road segments) and 22.82% (413 road segments) within the middle and outer rings respectively. Overall, 25.7% of the total road segments returns medium impedance level.

Low level of impedance is accounted for by 5.7%, 13.73% and 16% located within the inner, middle and outer ring. Low level impedance is accounted for by 11.8% (642 road segments). Out of these, 103 road segments, 248 road segments and 291 road segments are within the inner, middle and outer rings.

17.6% representing 956 road segments are categorised as very low impedance level. These road segments cuts across all of Metropolitan Melbourne with many of the roads within the outer ring. 3.5%, 8.52% and 40.8% are located within the inner, middle and outer rings respectively representing 63 road segments, 154 road segments and 739 road segments. Road segments within the outer rings account for over one third of very low impedance road segments.

126

All councils within the inner ring returns very high mean impedance level and are ranked 1 - 5. The councils return mean impedance level of 0.86 to 0.954. Melbourne City with a mean impedance level of 0.954 has the highest level of impedance followed by Yarra City which is ranked second (Figure 7).

Councils within the middle rings return different levels of medium and low impedance level. Out of the 17 councils within the ring, five cities return a mean impedance of between within the middle rings have a mean impedance level of between 0.577 – 0.800 with eight councils with medium impedance returning mean impedance level of 0.513 - 0.576. Three councils with 0.451 – 0.512 mean impedance level represent councils of low impedance. Banyule, Bayside and Moreland are councils with low mean impedance level.

Within the outer ring, Casey City returns a mean impedance level of 0.586 and falls within high impedance level category, while three councils, Cardinia, Hume and Wyndham Cities return medium mean impedance. The remaining councils within the outer rings fall within area of very low impedance level with Nillumbik returning the lowest impedance mean of 0.40.

Table 5. 5: Level of impedance comparison across the inner, middle and outer rings

Impedance Level % of Road Segments % of Road Segments % of Road Segments Inner Ring Middle Ring Outer Ring Very High impedance 48.7% 27.87% 7.62% High impedance 18.3% 19.38% 12.64% Medium impedance 23.76% 30.5% 22.82% Low impedance 5.7% 13.73% 16.06% Very Low impedance 3.5% 8.52% 40.86%

127

Inner Ring Middle Ring Outer Ring

48.70%

40.86%

30.50%

27.87%

23.76% 22.82%

19.38% 18.30%

16.06% 13.73%

12.64%

8.52%

7.62% 5.70%

3.50%

V E R Y H I G H HIGH MEDIUM LOW V E R Y L O W Figure 5. 7: Inner, middle and outer rings impedance level comparison. Figure 5.8: LMD impedance Levels across inner, middle and outer rings.

With increased urban density, the costs and lead times of last-mile delivery would increase (Boyer et al., 2004). Mapped transportation networks in Melbourne exhibit different levels of potential transportation network impedance to last-mile delivery. Some links in the network impede last-mile delivery more than others. Whether this spatial variability in LMD impedance across different constraints of urban form is statistically significant is yet to be investigated. This is tested using the Analysis of Variance (ANOVA) which allows comparison between the means. Generally, ANOVA is used to determine whether there exists any significant statistical difference between the means of a group of data.

The results of ANOVA (Table 5.7) show significant mean impedance across different suburban rings. The table reveals a mean impedance index of 0.39, 0.54 and 0.55 across the inner, middle and outer rings. Inner city councils can be clearly distinguishable from the councils located in the middle and outer ring, but the mean difference between middle and outer ring seems to be relatively small.

128

Table 5.6: ANOVA table across Metropolitan Melbourne

Constraints Mean Impedance Index Significance Population Density Low population density 0.52 Medium population density 0.88 High population density 1.00 Significance 0.00 Proximity to activity Centre Within 0.70 Outside 0.55 Significance 0.00 Railway boomgate Presence 0.54 Absence 0.58 Significance 0.00 Number of lanes 1-2 0.55 3 - 4 0.50 5 - 9 0.69 Significance 0.00 Speed limit 39 – 50 0.54 60 - 80 0.50 90 - 110 0.64 Significance 0.00 Land use Zone Residential Zone 0.54 Commercial Zone 0.47 Industrial Zone 0.50 Significance 0.00 Traffic Counts 0 – 3253 0.50 3254 – 7567 0.49 7568 – 11556 0.50 11567 - 19452 0.55 19453 - 124001 0.57 Significance 0.00 Truck counts 0 – 490 0.50 491 – 1100 0.52 1101 – 2400 0.56 2401 – 4900 0.66 4901 – 12001 0.60 Significance 0.00 Toll road With 0.34 Without 0.58 Significance 0.00

129

Metropolitan Wide Inner 0.3949 Middle 0.5471 Outer 0.5569 Significance 0.00

The results of ANOVA are discussed by individually linking with the key constraints of urban environment including built and regulatory.

5.3.1 Population Density

Generally, the need to control urban sprawl resulting from population increases within Metropolitan Melbourne necessitated the introduction and implementation of a compact city model. This is to achieve higher population density in areas situated within or in closer vicinity of the designated activity centres. The concentration therefore drives the need to cater for the dense population within the inner city through transport strategies and planning policies like the creation of principal and major activity centres across the Metropolitan Melbourne. As a result, roads get narrower through creation of bus lanes and bicycle path and more congested, with parking spaces getting harder to find, especially for last mile delivery. Councils within the inner ring and activity centres, high population density records high level of impedance with higher average, while the outer ring with low population density and areas farther from activity centres returns low impedance average.

Areas with higher population density especially within the inner rings account for very high level of impedance, while areas with low population density fall within areas of very low impedance. Port Phillips, Stonnington and Melbourne within the inner rings return very high level of impedance, especially within area of proximity to principal and major activity centres. Road segments with very high impedance represent 10.4% (567 segments). Yarra Ranges, and Melton with a population densities of 1:56, and 1:98 are areas of very low impedance. Segments within the very low impedance category represent 18.2% with 987 road segments.

ANOVA test reveals a significance relationship between impedance levels and population density. Mean impedance index for population reveals 0.52, 0.88 and 1 within area of low, medium and high population densities respectively (Table 5.7). Area of low population can be clearly distinguishable from areas of medium and high population densities. However, the

130

mean difference between the medium and high population densities appears to be relatively small, 0.88 and 1.0 respectively.

5.3.2 Proximity to Activity Centres

Areas recently zoned Activity Centre (either principal and major activity centres) within Melbourne Metropolitan Council in accordance with the policy direction of Clause 37.08 of the Victoria Planning Provision will continue to experience high population growth and high level of impedance index in accordance with the population density increase within these activity centres (See Clause 37.08 of the Victoria Planning Provision).

There is a significant and positive relationship between LMD impedance and proximity to activity centres. Areas with closer proximity to activity centres returns higher level of impedance compared to areas farther from activity centres. The ANOVA results reveal a significant difference with respect to proximity to activity centres. There is a clear difference between the mean impedance of areas within a distance of 1 kilometre from the activity centres. There appears a large mean difference between area of close proximity to activity centres with mean impedance of 0.70 and areas farther away from activity centres returning a mean impedance of 0.55.

5.3.3 Tollways

All tollways are located within the inner and middle rings. These networks of roads are characterised by higher speed limits ranging between 80km/h and 100km/h and 3 or more lanes on each direction. While roads with toll are to make delivery faster, drivers of small- scale delivery are however frustrated with the dollar amount they have to pay to use such toll roads. When used, it costs between $14.88 to $28.45 and $9.99 and $10.29 to travel between Tullamarine to Monash Freeway (Toorak) on City Lin and East line respectively for Light Commercial Vehicle (LCV) and Heavy commercial Vehicle (HCV) last mile delivery vehicles. In that sense, toll roads are negative constraint on last mile delivery for those who intend to avoid them.

The use of toll way between Tullamarine Freeway and Monash Freeway can result in approximately 20 minutes shorter travel time and avoidance of 39 traffic lights as well as a reduction of Co2 contribution from 4.5kg to 5.7 kg. In terms of fuel consumption, it is a

131

reduction from approximately 2.8 litres. In monetary terms, 20 minutes increased delivery time compared to approximately $3.64 (if fuel cost $1.3 per litre) is not considered a significant to last mile delivery vehicle drivers especially during peak period of travel.

Impedance index for road segments with toll and without toll return a low mean impedance of 0.34 compared to segments with no toll with a mean index of 0.58. This reveals a clear distinguishable mean difference between road segments with toll and without toll. The difference could be driven by higher speed limits, no road intersections and other road-related constraints.

5.3.4 Landuse Zoning

Landuse zoning as a planning control presents a different level of impedance to last mile delivery. The type of zone influences the hierarchy of Roads (Eppel et al., 2001). The hierarchy of road subsequently determine the speed limit. Specifically, residential zones and areas around activity centres are often assigned lower speed limits compared with other zones. The relative lower speed in these zones affects delivery time. Road constraints such as reduced number of lanes, increased number of traffic lights, prevalence of bus lanes and bus infrastructure and tram routes also characterise such residential zones and tend to reduce the efficiency of last mile delivery within residential zones.

Areas within industrial zone and road zone (outside the residential zone and activity centre zone) provide opportunity for higher speed limit, less traffic lights and availability of shoulder for parking for last mile delivery. There appears no clear distinguishable difference between the impedance means of residential, commercial and industrial zones as Table 5.7 reveals mean impedance difference of 0.54, 0.47 and 0.50.

5.3.5 Speed limits

Speed limit affects the distance that can be covered within specific time period. The higher the speed limit, the faster last mile delivery vehicles can travel. Areas of higher speed limit especially with 80 km/h to 100 km/h will not be within close proximity school zones, will not have railway boom gates/rail crossing and tram routes. These areas outside the residential zone or activity centre zone will have lower impedance level on last mile delivery and are common within the outer ring. It is assumed that the lower the speed limit, the higher the

132

level of impedance. There appears no clear difference between areas of 39 – 50 km/h and 60 – 80 km/h. This may be due to other constraints such as traffic counts and zoning associated with such roads. However, there are clear differences between 39 – 50km/h and 90 – 110km/h.

5.3.6 Trams and Railway Boomgate

Availability and proximity of tram lanes are common within the inner and middle rings. Tramways shape the geography of and add character to Melbourne. There are however delays associated with same grade movements of trams and vehicles Last mile delivery vehicles parked for pick-up or delivery in areas where there is a tram route will often be illegal in Melbourne CBD. Such parking prevents other vehicle through pass and can result in fines. As shown in Figure 5.8, vehicles will have difficulty getting a through-pass as a result of the deliery van parked on the road shared with a tram route.

Figure 5.9: Last mile delivery and tram route interface (Source: Google maps)

Boomgates are significant constraint on vehicle movement in general and on last mile delivery in particular. Where last mile delivery vehicles want to avoid boomgate closure in some cases, existence of low height bridges prevents last mile delivery vehicles from using such underpass where boomgates can be avoided (Figures 5.9 and 5.10). For example the low height underpass along Cheltenham Road in Dandenong City forces last mile delivery

133

vehicles travelling to Princess Highway to go through Hammond Road and Webster Street. Such vehicles going through Webster Street to the Princess Highway are likely to encounter delays due to railway boomgate along Webster Street. In addition, the use of some railway underpass by last mile delivery vehicles result in collision with the underpass (Figure 5.9). The removal of boomgates through the Victoria level crossings will help reduce the friction of this constraint on the speed of LMD vehicles. The rail overpass resulting from the removal of the boomgates are high enough to allow last mile delivery vehicles without any collision.

Figure 5.10: Height limit constraint along Cheltenham Road in Dandenong (Source: Google maps)

Figure 5.11: Webster Street Dandenong railway boomgate (Source: Google maps)

134

Figure 5.12: Overhead collision resulting from low height bridge

5.4 Local Area Analysis

From a council planning perspective, it is important to dissect the analysis at a local area level. In this study, three council areas are selected to provide further analysis at a council level. Problems are readily understandable and potential solutions are practically more implementable by local authority. The cases considered in this study include Maribyrnong (inner ring), Monash (middle ring) and Nillumbik (outer ring). These councils also represent different planning contexts, socio-cultural geographies and economic profiles. Table 5.8 highlights the key differences across these councils. Maribyrnong is high-population density inner city council with higher LMD impedance on dense road network. Maribyrnong City, given its location, is also a gateway to Metropolitan Melbourne as it provides direct access to the Port of Melbourne. Maribyrnong City returns an impedance level of 0.864 and classified as very high impedance.

Monash City located within the middle rings also has high population density, slightly lower than Maribyrnong. Transport network density within the council is less than that of Maribyrnong. Monash Freeway traversing through the council can be attributed to the medium level of impedance of 0.55 experienced by the council.

135

Nillumbik Shire on the other hand is a shire council located within the outer ring of Metropolitan Melbourne. The very low impedance return of the council can be attributed to lower population density and lesser number of road segments within the council. Such road segments have less traffic lights, no railway boomgates or tram routes and also experience higher speed limit.

Table 5. 7: Small area analysis

Attributes Maribyrnong Monash Nillumbik Population Density per square km 1916 1891 133 Distance of Council office from CBD 9 23 25 (kilometre) Number of Road Segments 134 343 102 Mean Impedance 0.864 0.55 0.41 Rank 3 14 31

The breakdown of impedance levels of the selected councils are presented in Table 5.9 and Figure 5.13. 21%, 12% and 56% of road segments within Maribyrnong, Monash and Nillumbik return higher levels of LMD impedance respectively.

Lower impedance levels are witnessed within Nillumbik compared to Maribyrnong and Monash. Overall, impedance levels within Nillumbik appear as a reverse of impedance level of overall Metropolitan Melbourne. For example, Nillumbik returns 7%, 9%, 27.8%, 24% and 34.16% from very high to very low level of impedance compared to Metropolitan Melbourne with 28.1% 16.8%, 25.7%, 11.8% and 17.6% respectively from very high to very low.

Table 5. 8: Metropolitan Melbourne and small area level of impedance

Level of Melbourne No. of Maribyrnong No. of Monash No. of Nillumbik No. of Impedance Metro Road Road Road Road Segments Segments Segments Segments Very High 28.1 1522 18% 24 15.3% 59 5% 7 High 16.8 911 12% 16 13% 38 9% 12 Medium 25.7 1394 45% 60 43.7% 150 27.8% 37 Low 11.8 642 8% 11 13% 45 24.04% 32 Very Low 17.6 956 17% 23 15% 51 34.16% 45

136

50%

45%

40%

35%

30%

25%

20%

15%

10%

0% Very High High Impedance Medium Impdeance Low Impedance Very Low Impedance Impedance

Melbourne Metro Maribyrnong City Monash City Nillumbik City

Figure 5.13: LMD impedance levels across different councils

While Metropolitan Melbourne returns more very high impedance levels, Monash displays an overall medium impedance level. Maribyrnong however, returns significant high-to-medium impedance compared to Metropolitan Melbourne. Overall, road segments with very low impedance level are more in the three councils compared to the overall Metropolitan Melbourne. Within the Nillumbik Shire Council, the proportion of road segments with low and very low accounts for over 50% with only 14% road segments falling into high to very high category. This is in line with the overall population density of 133 person per square kilometre, compared to Maribyrnong and Monash with 1916 and 1891 person per square kilometre respectively.

Specifically, the analyses of last mile delivery impedance index revealed significant differences in transportation network impedance levels across Maribyrnong City. This spatial variability in LMD impedance across these three councils indicates the differences in network capacity and planning controls imposed through land use regulations and road infrastructure. Overall, a higher percentage of the road network within Maribyrnong presents medium impedance level. Network with low impedance remained restricted to areas with lesser number of intersections, and traffic lights.

Within the Maribyrnong City, higher levels of impedance are registered in the Footscray Central Activity Centre, which represents high density living and mixed land use. In particular,

137

the northern part of the centre, which is currently experiencing high population growth, has returned high impedance scores. This is despite the fact that the proposed population density increase within the activity centre in accordance with the policy direction of Clause 37.08 of the Maribyrnong City Planning Scheme is yet to take full effect.

Overall, Gordon Street, Ballarat Road, Raleigh Road, Francis Road and Whitehall Road, Sunshine Road, and Ashley Street are of high impedance levels (Figure 5.14), as they attract significant number of heavy trucks that operates along these routes. Gordon Street is further congested because of tram lanes that run along the street, connecting Footscray Activity Centre to Highpoint Shopping Centre. Specific last-mile delivery strategies including clearway and delivery corridors to address movements of trucks within the city are lacking.

Figure 5.14: LMD impedance levels within Maribyrnong City.

In addition, the areas are in proximity with the Port of Melbourne and warehouses serving the western suburb of Melbourne. These road networks traverse through residential developments with reduced speed limit, prevalence of railway crossings and overhead bridges (from train tracks), truck restrictions, higher traffic light density and bus lanes and lacks of service lanes. Specifically, Sunshine Road and Ashley Street are characterised by single lane width, and traffic lights, which impede the delivery of goods within the inner city locations (Figures 5.12).

138

In contrast, the roads segments that hold medium and low LMD impedance scores are characterised by fewer road intersections and traffic lights. In addition, these routes do not have major shopping strips, tram lanes and bus lanes/stops.

5.5 Summary

This chapter identified the key LMD constraints, computed LMD impedance index and mapped the spatial variability of LMD impedance on the road network in Melbourne. These constraints relate to the built environment, transport network and planning systems controls. Analyses were carried out both at metropolitan and local council area levels.

LMD impedance levels vary across Metropolitan Melbourne and at a local level. High impedance levels are found to exist within inner ring of Metropolitan Melbourne and along major arterial roads. Also, impedance levels tend to reduce with increasing distance from Melbourne CBD. ANOVA results showed statistically significant relationships between LMD and constraints. Road segments with high impedance are likely to be characterised by high population density, few lanes, lower speed limit, high number of traffic lights and within 1 kilometre proximity to Activity Centres with direct conflict with tram routes. Road segments with low impedance on the other hand are characterised by lower population density, more lanes, reduced number of traffic lights and no direct conflicts with tram routes or bus lanes.

The next chapter proceeds with discussion on stages 2 – 5 of the research design. This includes clustering and defining the constraints into dimensions, generating the impact and uncertainty matrix, as well as building LMD scenarios and carrying discussion on stakeholders analysis.

139

CHAPTER SIX

BUILDING LAST MILE DELIVERY SCENARIOS

6.1 Introduction

This chapter purports to formulate future LMD scenarios by implementing the remaining stages of Scenario Thinking method. Extracting higher-order dimensions that collectively represent the underlying structure behind 34 constraints is pivotal to building a better understanding of the key drivers of LMD. The severity and uncertainty of impact of these dimensions are vital input in shaping future plausible scenarios with different outcomes. Stakeholders with different levels of power and interest to cause change in LMD are also discussed in the chapter. The chapter discusses what to change to and how as well as who to cause change in LMD.

This chapter is organised into seven sections. Section 6.2 discusses the overall framework for building LMD scenarios, while section 6.3 identifies the underlying dimensions generated through the clustering process. Section 6.4 uses the clustered dimensions to generate an impact and uncertainty matrix. The different future scenarios will be formulated in section 6.5 to answer the research question 3 what to change to? Section 6.6 undertakes a stakeholder analysis using a two-dimensional framework (Interest and Power) to identify key players/actors that could cause the change required to improve LMD efficiency in Melbourne. The summary of the chapter is presented in section 6.7.

6.2 Building Last Mile Delivery Scenarios

In Chapter 5, stage 1 of the Scenario Thinking was implemented to estimate and map last mile delivery impedance. This chapter proceeds with the remaining stages with the aim of building possible and plausible LMD scenarios (see Figure 6.1). The chapter continues the discussion of ‘what to change’, ‘what to change to’ and ‘how to cause change’. Furthermore, the question to be answered in this chapter is about ‘what if?” in terms of last mile delivery in Metropolitan Melbourne. What if the population continues to grow and the government continues to implement policies and regulations that do not favour last mile delivery? What if the road continues to get narrower and last mile delivery vehicles are not able to make just in

140

time delivery as a result of decayed infrastructure? What if infrastructure continues not to favour last mile delivery or fails to match last mile delivery?

Last mile delivery scenarios represent the strategic knowledge-sharing and conversation to develop possible and plausible futures. Deliberately, building the scenarios seeks to envisage the future of last mile delivery in the context of changing the built and regulatory environment. The consideration of the present and trends of development of cities will provide a better understanding of multiple futures of last mile logistics.

Stage 1 of Scenario Thinking workshop has enabled the identification and mapping of LMD impedance index; whilst the remaining stages are discussed in this chapter. As shown in Figure 6.1, these stages include clustering of constraints into meaningful few dimensions of utmost importance, generation of impact and uncertainty matrix, scoping and development of scenarios and stakeholder analysis.

Stage 2: Clustering of constraints and defining of clusters

Stage 3: Generating of Impact and Uncertainty matrix

Stage 4: Scoping and Development of LMD Scenarios

Stage 5: Stakeholder Power and Interest Matrix

Localised last mile delivery mitigation strategy

Figure 6.1: Research Design Stages

6.3 Stage Two: Clustering of Last Mile Delivery Constraints

Clustering of LMD constraints is stage 2 of the Scenario Thinking process. In all, 34 built environment, planning and transport systems constraints were identified in the workshop that impede the efficiency of last mile delivery.

141

The aim of the clustering exercise is to find one set (out of an indefinite number of possibilities) of similarities through commonalities of the 34 constraints. This is to differentiate the groups from each other through commonalities (Ringland, 2006). Smaller number of higher order dimension(s) of similar constraints emerges through clustering.

The identified constraints are clustered through a group process of action discussion and consensus building. Constraints are clustered by ensuring coherence within each cluster. Constraints in each cluster can have either a positive or negative effect on last-mile delivery (Table 5.1). Hence constraints can aid or impede last mile delivery. For example, lower speed limits hinder (negative) the movement of freight vehicles; while higher speed limits reduce lead time (positive) by expediting vehicle movement. Therefore, exploiting the constraints requires maximising positive effects of constraints; whilst minimising negative effects on last mile delivery. Subordinating requires the reduction process to eliminate negative effects of constraints. Workshop participants were requested to collectively cluster the identified constraints into six dimensions. The collapsed dimensions reflect the “higher order dimensions”, which represent the constraints of built environment, planning and transport systems. The six dimensions that emerged and agreed to by all participants and were labelled by participants:

(i) Freight Infrastructure,

(ii) Landuse Intensity,

(iii) Infrastructure Supply,

(iv) Infrastructure Sharing,

(v) Intersection Control,

(vi) Human Behaviour.

6.3.1 Freight Infrastructure

The constraints clustered into the Freight Infrastructure dimension relates to the basic physical infrastructure and facilities required for provision of last mile logistics (Figure 6.2). Three constraints are clustered into this dimension. These constraints include loading and unloading areas, number of available loading and unloading areas and parking restrictions. These constraints are logistics infrastructure in cities that enables pickup and drop off of goods. As well, loading and unloading can be on-street as part of the road network or off-

142

street. Specifically, availability of loading and unloading areas will assist in quick loading and unloading turnaround, reduced lead-time and cost associated from delays. Parking restrictions on the other hand limit availability of spaces for loading and unloading of freight vehicles. Where parking restrictions are not available, it can result into illegal parking and infringement notice, which in turn adds to the cost of last mile delivery.

Loading and unloading areas

Number of available loading Freight Infrastructure and unloading areas

Parking restrictions

Figure 6.2: Key constraints within the Freight Infrastructure dimension

Loading and unloading areas are restricted space for use by delivery vehicles and are often designated with a sign to indicate the exclusiveness for loading and unloading functions. Vicroads (2000) specifically listed the following vehicles to use on-street loading areas.

 Delivery vehicles displaying the company business name on each side  Courier vehicles appropriately signed  Trucks  Public buses and taxis  Other commercial passenger vehicles.

From this list, while other vehicles can use an on-street loading zone, delivery vehicles are proscribed from accessing taxi ranks and bus stops. In addition, private cars deprive last mile delivery vehicles opportunities to park in designated loading zones (Figure 6.3). A delivery vehicle will therefore, be deprived of a loading zone parked on by other vehicles. Prior to 18 January 2018, the Victoria Planning Provision prescribed the requirements of a loading area as a part of an industrial development.

143

Figure 6.3: Private cars parking within loading zone

When provided off-street loading facility, manoeuvring of vehicles in and out of the loading zone from the street is often frustrating especially if a vehicle parks opposite the entry of the access of the loading zone.

6.3.2 Landuse Intensity

Landuse intensity is the magnitude to which a parcel of land is being used or developed in conformity with zoning ordinances. In economic terms, the interaction between the values determined by location based on land uses and rent payable on the land determines the intensity of the land. The classical concentric model of Von Thunen (1826) applies in the level of landuse intensity and value from the city centre to the hinterland.

Planning systems use zoning as a measure of control of the extent of use of any parcel of land. The intensity could be high, medium or low depending on the zoning provision. The zoning provision on Victoria is controlled via the different Municipal Planning Scheme. Broadly, different zones within the Victoria planning provisions include residential, commercial, industrial, and activity centres zones. The land use zoning as a driving force combines with other road infrastructure related constraints like speed limit, while provision of loading and unloading areas impacts last mile delivery vehicles differently.

Eight constraints on last mile delivery are clustered into this dimension. Constraints within the dimension of Landuse Intensity relates to urban planning and zoning. The dimension includes surrounding area and vehicle ownership, traffic count, dwelling density, size of shopping centres, traffic generation and activity centres (Figure 6.4). This dimension is

144

largely associated with population density and vehicle ownership as well as the status of shopping centres and their locations. Larger regional shopping centres are more likely to attract more consumers and thus require larger number of freight vehicles. The size of the shopping centres is also a function of the applicable zoning that dictates their sizes and functions.

Planning provisions utilise the zoning of surrounding areas to restrict the size of a shopping centre. Also, planning provision under Clause 52.06 of the Victoria Planning Provision (VPP) requires reduction in car parking or non-provision within the surrounding area within closer proximity to public transport. In such areas, residents are encouraged to use public transport, which intend to reduce vehicle ownership especially within the Activity Centres.

Surrounding area and vehicle ownership

Traffic Count

Dwelling/Population Density

Year of Subdivision

Landuse Intensity

CBD and Commercial landuse

Size of shopping Centres

Number of shops (traffic generation and parking

Activity Centre Proximity

Figure 6.4: Key constraints within the Landuse Intensity dimension

145

6.3.3 Infrastructure Supply

Infrastructure Supply relates to both physical and material services required by a logistics system to function properly. The adequacy or inadequacy of such Infrastructure Supply can promote or hinder the last mile delivery. In a transportation network, the road pavement and associated services, for instance, must be adequate to enhance last mile delivery due to the size and shape of freight vehicles used in LMD.

The dimension for Infrastructure Supply includes the largest number of constraints. The constraints loaded on this dimension are made up of intersection constraints, speed limit and change in speed limit, road network capacity, change in road hierarchy and change in number of lanes, road closure (Figure 6.5). Others are road width, number of lanes, distance from freight network, road network design and alignment (road geometry), road hierarchy and toll. The speed limit on a road is a function of the hierarchy of such roads. While speed on arterial roads can be up to 80 – 100km/h, street and collector roads are between 40 and 60km/h.

On any road network, intersection can either be at same grade or separated. Where at the same grade, three basic forms of intersection can be identified. The first one is the “T” intersection which is described as three leg intersection. The second and third being cross intersection and multi leg intersection. The controls on each intersection can be signalised or un-signalised. Each form of same grade intersection whether signalised or non-signalised has varying impedance levels on movement of freight fleet. Separated intersection favours traffic flows while delays result from same grade intersection. The association of intersection with road capacity, traffic lights, road hierarchy and alignment are important consideration, which participants considered as important in last mile delivery.

Also considered important are constraints on last mile delivery through road closure or prohibition of using some local roads even when such roads are the best connection to pick up or drop off location. Abrupt changes in the hierarchy of road and subsequent speed limit changes also restrict the use of road infrastructure by last mile delivery vehicles.

146

Intersection

Speed limit and change in speed limit,

Road network capacity

Change in road hierarchy and change in number of lanes

Road closure

Road width

Number of lanes

Distance from freight network Infrastructure Supply

Road network design & alignment

Road hierarchy

Toll

Abrupt cahnge in speed limit

Abrupt cahnge in number of lanes

Location of adjascent arterial road

Proximity to freeway interchange

Figure 6.5: Key constraints within the Infrastructure Supply dimension

147

6.3.4 Intersection Controls

The fourth dimension identified by workshop participants is intersection controls. Where a grade separation is not available, intersection controls are often adopted to control traffic to limit conflicts and crashes within the transport network. Participants acknowledged the advantage of roundabout as a better intersection control compared to traffic lights. Three constraints are clustered into this dimension, which includes railway boom gate, traffic lights and road design after thought (Figure 6.6). The intervals between traffic lights and associated timing along a stretch of road can have either positive or negative effect on the efficiency of last mile delivery vehicles. Shorter timing of traffic lights and closer intervals will result in delays and loss of valuable time.

Railway Boomgate

Intersection Controls Traffic lights

Road design afterthought

Figure 6.6: Key constraints within the Intersection Controls dimension

Traffic light is found to be one of the intersection controls that influence LMD. The constraint of traffic lights at an intersection astronomically increases with the number of conflict points at intersection. While conflict points can be as low as 6 in a three-leg intersection, it increases to 24 and 120 in a four leg and five leg intersections. Delays therefore increase with the number of legs and conflicts points (Mainroads Western Australia, 2015) at intersection.

The issue with road design afterthought is associated with delays during the construction as well as the geometry and alignment resulting from such afterthought design. An introduction of a roundabout on a road as afterthought can greatly alter the road usage. Cutting out of bicycle lanes and bus lanes also reduces number of lanes on a road and can affect manoeuvring along the road segment.

148

6.3.5 Infrastructure Sharing

Three constraints are clustered into the dimension of infrastructure sharing. These constraints are bicycle lanes, trams and proximity to other transport modes (Railway/Tram) as captured in Figure 6.7. Infrastructure sharing dimension relates to other modes of transport that share the road with last mile delivery vehicles especially when at the same grade level. For example, bicycle lanes carved out of the road as an afterthought are considered as limiting manoeuvring and slowing down last mile delivery vehicles. Compulsory stoppage behind trams at tram stop is recognised as a major constraint on last mile delivery vehicles. There are often no or limited opportunity for parking for loading and unloading on shared roads, which result in considerable walking distance for pickup or delivery.

Bicycle Lanes

Tram Lanes Infrastructure Sharing

Proximity to other land transport modes

Figure 6.7: Key constraints within the Infrastructure Sharing dimension

Land transport generally includes road and rail transportation, hence sharing the road with other modes of transportation is therefore inevitable. The road is often shared with same grade rail for tram and bicycles (Figure 6.8).

149

Figure 6.8: Typical bicycle lane cut out from existing road network (Source: Google maps)

6.3.6 Human Behaviour

The last dimension is Human Behaviour, which is characterised by two constraints; namely, the lack of understanding of road rules, and road wok restrictions when there are no road works (Figure 6.9). The constraints in this dimension are of high importance to drivers because of the size of their vehicles and frustration experienced by drivers when other road users take advantage of low take off speed of last mile delivery vehicles.

Lack of drivers understanding of road rules

Human Behaviour

Road work restrictions

Figure 6.9: Key constraints within the Human Behaviour dimension

While the case of drink driving cannot be categorical, speeding to merge in front of last mile delivery vehicles is acknowledged by participants as constraint which results in accidents and road rage in many times.

150

6.4 Stage Three: Generating Impact and Uncertainty Matrix

The next stage in Scenario Thinking method is the generation of an impact and uncertainty matrix. The purpose of generating the impact and uncertainty matrix is to understand the impacts of the underlying dimensions on last mile delivery and uncertainty associated with those impacts. Impact refers to the likely effect (i.e. high-low) of the identified dimensions on the efficiency and performance of last-mile delivery in cities: whilst certainty is the levels of confidence on the likelihood of the outcomes.

The six clustered higher-order dimensions namely ‘Freight Infrastructure’, ‘Landuse Intensity’, ‘Infrastructure Supply’, ‘Infrastructure Sharing’, ‘Intersection Control’, and ‘Human Behaviour’ listed on a Post-it® are arranged along the full length of the impact axis, positioning them relative to each other through a process of discussion of the perceived degree of impact/ impedance of each constraint on last mile delivery.

After the impact and levels of uncertainty are assessed on a two-dimensional space, the group proceeded to position the Post-it® where each dimension is relocated relative to the impact/uncertainty matrix, that is along High/Low Certainty (of the impact/impedance of the outcomes) axis. The two-dimension space representing impact and uncertainty is presented to stakeholders on a scale of low to high along the continuum. This is done using the full extent of the axis and without disturbing the relationship on the Impact axis. The positioning is to help ascertain the likely impact of each dimension on the focal issue and relative degree of uncertainty on the potential outcomes.

The two-dimensional framework driven by Infrastructure Supply and Landuse Intensity based on high impact on LMD and high uncertainty of the potential outcomes in the future were developed. Figure 6.10 shows the positioning of all dimensions on the level of certainty and impact.

Freight Infrastructure is identified to be of low impact and medium impedance and located within the Quadrant 1 of low impedance and high certainty. Two dimensions, Infrastructure Sharing and Intersection Controls are of moderately rated on impact and certainty and positioned within Quadrant 2. Human Behaviour is considered to be of low impedance and low certainty and positioned within Quadrant 3. Infrastructure Supply (A) and Landuse

151

Intensity (B) are rated to be of high impact and of high uncertainty and positioned within Quadrant 4.

The dimensions within Quadrant 4 are of high impact and high uncertainty as the participants are not sure of the likely outcomes. Hence they are of high impact of the constraints identified by the stakeholders and high uncertainty of the potential outcomes. These constraints are considered by participants as independent of each other and are used to form the last mile delivery impedance scenario dimensions. The dimensions are identified as A and B to be used for the building of future scenarios.

High Low Impedance Low Impedance High Certainty Q1 Q2 High Certainty

Infrastructure Freight Sharing Infrastructure

Intersection

Control UNCERTAINTY A Infrastructure Supply B Human Behaviour Landuse Intensity

Low Impedance High Impedance Q3 Q4 Low Low Certainty High Certainty High IMPACT Figure 6.10: Impact -Uncertainty matrix – higher order factor.

6.5 Stage Four: Scoping and Developing LMD Scenarios

Stage 4 purports to scope and develop LMD scenarios using the two key dimensions identified which reflect higher impact and higher uncertainty. This stage aligns with the first principle of TOC by answering the question: what to change to? Efficient delivery and reduction in cost of last mile logistics despite the constraints of the built and regulatory

152

environment are the goals. This is what building different scenarios aims at achieving. These two dimensions of Infrastructure Supply and Landuse Intensity are used to portray the future of last mile logistics.

Scoping of the future outcomes for last mile delivery is conducted through an in-depth discussion of the two dimensions towards answering the “what if” question. Answering this question aids in development of scenarios. What if the constraints in Landuse Intensity are at the worst or at their best and Infrastructure Supply at their worst or best? A consideration of the positive and negative features of the constraints is the focus of the scoping and building of the scenarios. For example, the participants have identified growing complexity of LMD resulting from the compact city model and discussed the following what-if situations prior to the development of LMD scenarios.

(i) increased/decreased vehicle ownerships resulting from higher density living without increased road capacity or use of other transportation modes like increased public transportation; (ii) more restrictive zoning with imposition of lower speed limits in inner city suburbs; (iii) a reduction or more loading opportunity for last mile delivery vehicles resulting from the removal of the provision of loading and unloading provisions from the VPP; (iv) decreased or increased speed limits; number of lanes/lane width; (v) increased or decreased distance from freight network; and (vi) increased or decreased or no tollway etc.

Using the two selected dimensions of Infrastructure Supply and Landuse Intensity in building the LMD scenarios, participants considered the future of last mile delivery over the next 10 years within Metropolitan Melbourne. A 10 year period is selected based on the planning horizon and projected urban development plans in Melbourne. While urban planning often uses a 10-year planning horizon (Bradfield et al., 2005), 72% of scenario thinking operators have used 10 year planning horizon (Boyomas et al., 2016). With the projected future impact and uncertainty of the dimensions in mind, participants discussed the broad possible and plausible scenario outcomes.

153

These reflections by the participants are represented in terms of best/worst, best/best, worst/worst and worst/best scenarios, which were constructed using the two dimensions (i.e. Infrastructure Supply and Landuse Intensity). The best/best outcomes describe an ‘ideal world’, which the dimensions favours last mile delivery. However, last mile impedance issues arise by consideration of the best/worst outcomes of the dimensions, a scenario in which one set of positive descriptors was moderated by a set of largely negative descriptors.

The four key scenarios developed are interpreted in four quadrants to provide the basis for the development of LMD scenarios on a best/worst interaction of the dimensions. The quadrants represent a set of well-defined four possible LMD outcomes. The matrix represented in Quadrant 1 is the best of Infrastructure Supply and worst of Landuse Intensity. Quadrant 2 represents best of both Infrastructure Supply and worst of Landuse Intensity, while quadrant 3 represents worst of both Infrastructure Supply and worst of Landuse Intensity. Quadrant 4 represents best of Landuse Intensity and worst of Infrastructure Supply. Infrastructure Supply (dimension A) on the vertical axis and Landuse Intensity (dimension B) on the horizontal axis define the anchors for formulating the future LMD scenarios.

The four scenarios formulated in the framework are not a prediction of the future, rather, an indication of the range of possible and plausible future outcomes under certain well- conceived conditions (Figure 6.11).

154

Best/Worst (Q1) Best/Best (Q2)

 Efficient Delivery  Increase in usage  Fast and On Time  Decrease in Delivery congestion  Cost Effective  Low cost delivery Delivery  Increased  Higher speed and

productivity modal split.

SUPPLY  Gridlock and  Higher Environmental decrease in usage Impact  Damaged Products

INFRASTRUCTURE INFRASTRUCTURE  Delay increase  Productivity Loss  Congestion  Less Delivery per day  Gridlock  Negative  Business Stagnation Environmental  Damaged Infrastructure Impact  Ageing Infrastructure Worst/Worst (Q3) Worst/Best (Q4)

LANDUSE INTENSITY Figure 6.11: Scenario scoping – broad descriptions of future scenarios

6.5.1 Best/Worst Scenario

In Quadrant 1, Infrastructure Supply is at its best, while Landuse Intensity is at its worst. An indication of improvement in the identified constraints associated with Infrastructure Supply and a worst case for constraints clustered in Landuse Intensity that is unfavourable landuse to LMD. Participants agreed that in these circumstances, there is a possibility of an efficient and cost-effective delivery as a result of the best outcomes of the larger infrastructure capacity (Figure 6.12).

The best of Infrastructure Supply conditions includes lower constraint on last mile delivery vehicles in terms of reduced road closure and height restriction improved road alignment, increased number of lanes and reduced or no cost for using toll roads for last mile delivery vehicles. Best of Landuse Intensity, includes a reduction in traffic competing with last mile delivery vehicles, especially within inner city suburbs, commercial areas, activity centre zones and shopping precincts. On the contrary, the worst of Infrastructure Supply and Landuse Intensity includes continued closure and restriction on last mile delivery vehicles, increased toll, reduced number of lanes and more intensity landuse with increased private

155

vehicles competing with last mile delivery vehicles within activity centres and shopping stripes.

Best/Worst (Q1) Best/Best (Q2)  Efficient Delivery  Fast and On Time Delivery  Cost Effective Delivery  Higher speed and

modal split.

SUPPLY INFRASTRUCTURE INFRASTRUCTURE

Worst/Worst (Q3) Worst/Best (Q4)

LANDUSE INTENSITY Figure 6.12: Best (A)/Worst (B) scenario

6.5.2 Best/Best Scenario

Within Quadrant 2, the constraints clustered into Factors A and B are at their best. That means the best of both Infrastructure Supply and Landuse Intensity. With this scenario, there will be more certainty of speed limits. For instance, on Melbourne roads, speed limits can be reduced from 100km/h to 60km//h or even to 40km/h within sort distance on a road segment. Increase number of lanes, minimum disruption from road closures, and increased width of lanes to conveniently accommodate last mile delivery trucks are part of the best/best scenario. This will result into increased usage of the available infrastructure, decreased congestion due to better infrastructure, low cost delivery resulting from lower impedance and increased productivity (Figure 6.13).

This scenario is what last mile delivery aims at and is ‘what last mile delivery needs to change to’ for support of a cost efficient, sustainable and effective last mile delivery.

156

Best/Worst (Q1) Best/Best (Q2)

 Increase in usage  Decrease in congestion  Low cost delivery  Increased

productivity

SUPPLY INFRASTRUCTURE INFRASTRUCTURE

Worst/Worst (Q3) Worst/Best (Q4)

LANDUSE INTENSITY Figure 6.13: Best (A)/Best (A) scenario.

6.5.3 Worst/Worst Scenario

Quadrant 3 represents a worst/worst situation of the two factors. This will result in a likely collapse of last mile delivery infrastructure and congested transport network due to increased Landuse Intensity. The projected outcomes are opposite to the scenario in Quadrant 2 resulting from gridlock and decrease in usage, delay increase, increased and higher cost, productivity loss, lesser number of deliveries per day and negative environmental impacts such as poor air quality, increased greenhouse gas emissions, increased noise from last mile delivery vehicles, public safety and unsustainable delivery system (Figure 6.14).

157

Best/Worst (Q1) Best/Best (Q2)

SUPPLY  Gridlock and decrease in usage

INFRASTRUCTURE INFRASTRUCTURE  Delay increase  Productivity Loss  Less Delivery per day  Negative Environmental Impact Worst/Worst (Q3) Worst/Best (Q4)

LANDUSE INTENSITY Figure 6.14: Worst (A)/Worst (B) scenario

6.5.4 Worst/Best Scenario

Quadrant 4 represents the best of Factor B and the worst of Factor A. Within this scenario, Infrastructure Supply will be at its worst with reduced speed limit for last mile delivery vehicles, more road closures and reduced number of lanes including reduced lane width associated with increased traffic competing with last-mile delivery vehicles, considerable increase in dwelling and population density within the CBD and activity centres as well as surrounding land and longer travel distance from main arterial roads passing through restricted planning zones or through to the CBD to shopping centres. Participants largely agreed to the following possible and plausible outcomes of an ageing and damaged transport infrastructure, which results in deleterious impact on environment, damaged products, higher congestion, frequent gridlock and business stagnation (Figure 6.15).

158

Best/Worst (Q1) Best/Best (Q2)

SUPPLY  Higher Environmental Impact

 Damaged Products INFRASTRUCTURE INFRASTRUCTURE  Congestion  Gridlock  Business Stagnation  Damaged Infrastructure Worst/Worst (Q3)  AgeingWorst/Best Infrastructure (Q4)

LANDUSE INTENSITY Figure 6.15: Worst (A)/Best (B) scenario

6.6: Stage Five: Stakeholder Analysis - Power and Interest

Stakeholder analysis identified the actors who have power and interest to cause change. Stakeholders in last mile delivery have different levels of power and interest. Interest represents the level of concern by the stakeholder that is having a stake, while power represents the level of authority to influence or implement change in last mile logistics. Power is also the capacity to influence behaviour, cause change or ability to restructure situations (Mendelow and Aubrey, 1981).

A matrix with two axes, namely power and interest, are generated to identify key stakeholders directly involved in the management and delivery of last mile provisions. This matrix is often categorised into four quadrants, which are characterised and labelled as Context Setters, Players, Crowds and Subjects (Wright and Cairns, 2011). The Context Setters represent the unaffected; the Players are the actors; the Crowds represent the unaffected bystanders; and the Subjects are bystanders affected by decisions on last mile delivery. The subject are affected either as a result of the cost of delivery or delay in freight delivery.

159

The workshop participants identified a range of agencies and individuals as stakeholders in last mile logistics. Overall, the Federal, State and Local Governments, Truck Association, NGO, Road Users, Vicroads, Lobbyist, Traders Association, Port Authority, Business Owners, Local Community and Drivers are identified stakeholders in last mile delivery within Metropolitan Melbourne with different levels of power and interest and positioned relatively within the 4 quadrants (Figure 6.16). The participants agreed that power over the focal issue and interest in the matter can change over time. For instance, the interest of a particular person in the local community, which is a customer expecting a delivery can change from crowd to players if that delivery is linked to the performance of her business. Also depending on the situation, a context setter can become a crowd or player. The change in the position of these stakeholders on the matrix will impact the future of last mile delivery.

High Q1 Q2 Federal Government High Power High Power Transurban Public State and Transport Local Victoria Government Vicroads Low Interest High Interest Truck Port Association Authority Lobbyist Q3 Q4

POWER Low Power Low Power Business NGO Owners Low Interest Other Road High Interest Users Local Community Drivers Low High

INTEREST Figure 6.16: Stakeholder matrix – stakeholders to cause change in last mile delivery

6.6.1 Quadrant 1 - High Power-Low Interest

From the analysis, Administrators (i.e. Federal, State and Local Governments), Transurban and Trucks Association fall within Quadrant 1. They are identified as context setters because

160

they are set the polies and regulations that affects last mile delivery, however, they are unaffected. Stakeholders in this quadrant are recognised by the participants to have high power, but with low interest in the functioning of LMD. However, these agencies are responsible for regulations and policies to plan transport infrastructure and control city logistics provisions. The federal government sets general road rules at the national level; whilst they are implemented by state and local government. The Australian Road Rules are published by the National Road Transport Commission, which provides uniform road rules across Australia. It also includes appropriate definitions of offences and penalties for such offences.

Strategically, government as administrator can introduce land use zoning and corridors to facilitate faster last mile delivery vehicles movement and incorporate such into the relevant planning statutory documents. In addition, greater importance to loading and unloading especially in areas with close proximity to activity centres and apartment building in statutory planning assessment is important. Provision of such lies in the powers of state and local councils.

The power of Transurban lies in the setting of tolls. The toll charged on the roads and relative dollar and time fail to incentivise the use the tollway especially for sole business operators. Overtime, tolls charged on last mile delivery vehicles is between $14.88 and $27.90 for LCV and HCV respectively from Tullamarine to Toorak on Monash Freeway and $9.99 and $10.29 along the Eastern Freeway. Such operators avoid toll road and this leads to increased delay in last mile delivery.

The roads with tolls (Monash, Burnley Tunnel, and the East link, Bolte Bridge) are designed to reduce travel time. Such roads allow higher speed as well as no traffic lights or zero delays resulting from built environment characteristics. These roads normally bypass the congested low speed, shared/tram routes and boom gates that result in last mile delivery delay. However, smaller last mile delivery operators avoid such roads. Toll is considered as additional cost. The advantage of paying toll to the operators does not outweigh the time and cost of avoiding the roads. Reduction in toll paid by last mile delivery vehicles is in the power of Transurban. However, Transurban has low interest with high power. It however will not be interested in mitigating last mile delivery.

161

Participants positioned Transurban as having the lowest interest among all stakeholders followed by Truck Association. The federal government is positioned with the highest power of all the stakeholders.

The Australia Truck Association includes major logistics companies and transport industry associations in Australia. The association was established after the Grafton truck and bus crash in October 1989. One key task of the association is to develop national policies in conjunction with its member and lobby government to put such policies into effect. Participants identified the Association as having high power, but with low interest in last mile delivery. None of the drivers and logistic managers at the Scenario Thinking workshop identified any important role played by the Truck Association that could cause any positive change in last mile delivery.

6.6.2 Quadrant 2 - High Power-High Interest

This quadrant represents the stakeholders with high power and high interest. These are actors who have the potential to cause change in last mile logistics. Within this quadrant are Public Transport Victoria (PTV), Vicroads, Traders Association and Port Authority.

PTV is the statutory body established by the Victoria government to provide, coordinate and promote public transportation in Victoria. While the agency is statutorily public transportation, participants considered it to have high interest and power which can positively influence last mile logistics. Also, Vicroads has the key role in providing safe and easy connections to Victorians. Vicroads seek to “assist economic and regional development by managing and improving the effectiveness and efficiency of the road transport system as well as develop a more integrated and sustainable road transport system” (Vicroads, 2000). Generally, these two agencies are in charge of transport networks in Victoria. The interest of PTV and Vicroads lies within the transport network with more emphasis on public transport and regulating the movements of trucks.

Public Transport Victoria and Vicroads for example can influence speed limit and traffic lights management. Also, the removal of railway boom gates as currently being carried out and removal of other last mile delivery constrains are in their powers.

162

Each shopping centre and markets across Metropolitan Melbourne has its own trader association. These associations have high power and high interest in last mile logistics. At a local level, they lobby local councils. Traders Association as local lobbyist on the other hand mount pressure on local governments in terms of provision of vehicles loading and unloading for use of business premises. Despite the high power of the stakeholders in this quadrant, their power appears limited to register any impact on the constraints of last mile delivery as they cannot make any policy or legislation to improve last mile delivery. This is also despite their high interest.

6.6.3 Quadrant 3 - Low Power-Low Interest

Non-Governmental Organisations and other road users are located within this quadrant. These are categorised as Crowds as they represent unaffected bystanders in last mile logistics. Their power to effect change and interest in last mile delivery is relatively low. This group of stakeholders has no vested interest in last mile delivery. They are only interested in on-time delivery of goods.

With high level of car ownership in Melbourne, other road-users frustrate effective last mile delivery as a result of their actions. Particularly, LMD vehicle drivers are frustrated by lack of proper knowledge of lane changing exhibited by other road users. Such lane change and merging in front of last mile delivery vehicles sometimes results in collision between last mile delivery vehicles and other road users.

6.6.4 Quadrant 4 - Low Power-High Interest

Drivers, Local Community and Business owners are classified as Subjects. However, these stakeholders have low power but high interest in the last mile logistics. They are however directly affected by decision of the context setters and players. Majority of the drivers are identified as small scale operators, often on short-term contract. They are more likely to bear additional costs of last mile delivery and are most affected by last mile impedance. The extent of passing the cost to the end-users is often limited given the uncertainty of daily traffic flow.

163

Business Owners and Local Community (Endusers in B2B) are at the receiving end of the supply chain. Delays caused by the identified constraints increased delays and damaged products. Particularly, retailers lost income as a result of delays in delivery of products.

6.7 Summary

This chapter presented the analysis and results of the Scenario Thinking workshop and responded to the ‘what to change?’, ‘what to change to?’ and ‘how to cause change?’ These are demonstrated in stages 2, 4 and 5 of the research design. In all, 34 constraints were clustered into six higher-order dimensions which were driven through consensus building in stage 2. These six dimensions were (i) Freight Infrastructure, (ii) Landuse Intensity, (iii) Infrastructure Supply, (iv) Infrastructure Sharing, (v) Intersection Control, and (vi) Human Behaviour. Landuse Intensity and Infrastructure Supply represents the high impact/impedance and high uncertainty. These are the key dimensions that will drive the future of LMD.

Four scenarios relating to the future challenges of last mile delivery were created from the interplay of the Landuse Intensity and Infrastructure Supply dimensions. Stakeholders with power and interest were also identified who have the potential to cause change in last mile logistics provisions. The findings from the different quadrants are the likely outcomes of the projected future last mile delivery scenarios.

The chapter provided answers to research questions 2 and 3 through seamless integration of the principles of Theory of Constraints and Scenario Thinking. Chapter 7 formulates strategies from the scenarios towards mitigating LMD problems.

164

CHAPTER SEVEN

STRATEGIC FRAMEWORK FOR LAST MILE DELIVERY

7.1 Introduction

This chapter develops a strategic framework to mitigate the potential last mile delivery impedance. These strategies will be formulated from the outcomes of the LMD future scenarios. They relate to developing mechanisms and structures to tackle the challenges associated with projected future scenario-driven by the key question of “what to change to”.

This chapter is organised into five sections. Section 7.2 establishes the basis for the development of a strategic framework to respond to the challenges identified through the plausible LMD outcomes. The framework consists of key goals for last mile logistics and a set of desirable distribution objectives which the formulated strategies aim to achieve. Section 7.3 identifies the five key strategies, followed by Section 7.4 that discusses a range of actions recommended to help implement the strategies. Section 7.5 presents the summary of the chapter.

7.2 Development of a Strategic Framework for LMD

A strategic framework to mitigate last mile delivery challenges and associated problems resulting from emerging compact city is developed in this section.

A strategic framework is a well-thought-out process developed to build initiatives and support key objectives and future-oriented actions (Mulcaster, 2008). It seeks to develop policy actions to ensure a phenomenon performs efficiently and effectively at high standards. It is a combination of a guiding plan, strategies and actions to achieve LMD objectives. The strategies formulated in this thesis are integrated in perspective because it incorporates different actions to achieve key desirable objectives and ultimately, the goal of LMD.

165

The goal of the developed last mile delivery strategies is to achieve an efficient, sustainable and low-cost delivery of goods in large cities. The goals are broken down into attainable objectives, which would have to be achieved via possible actions.

Figure 7.1 illustrates the four integrated levels of the developed LMD strategic framework and the links between the levels (Rumelt, 2011). The framework links the goals and objectives with actions through different strategies.

Goal

Objectives

Strategies

Actions

Figure 7.1: A Strategic Framework for an efficient, sustainable and cost effective LMD

To develop this framework, the scenario outcomes are first translated into six major objectives, which are connected to the key goal of last mile delivery. Five strategies along with a combination of actions are developed to achieve these distribution objectives (Figure 7.2).

166

Strategies Actions

LMD Objectives Adaptive vehicles and Land-use zoning strategy delivery time Reduce gridlock and congestion associated with last mile delivery Consolidation and load factor efficiency Remove last mile delivery Last-mile corridor delay strategy Urban Consolidation Centre

Increase last mile delivery productivity and reduce cost Shared economy model Distribution network strategy Underground loading Increased daily LMD and unloading area

Multi modal Delivery tunnel Reduced or zero environmental impact transportation strategy

Collaboration Increase LMD usage of transport network system Stakeholder Engagement Intelligent strategy Transportation Systems

Figure 7.2: Strategies, key desirable distribution objectives and applicable actions.

7.2.1 Setting the Last Mile Delivery Goals

The first step of any strategic framework is to set goals. Goals are future desirable outcomes to which all efforts are directed. This can be short-term or long-term. The uncertainty associated with the future often necessitates setting up a desirable future, which the society aspires for. Setting goals therefore assists in the establishment of a clear direction (Grant, 2003) or pathway.

The strategic goal formulated for last mile delivery is to ensure an efficient, sustainable and low-cost delivery to enhance city liveability and urban economy. This goal is formulated given the current challenges, and future desirable outcomes and uncertainty. Current literature indicates that last mile delivery is the most inefficient and costliest component of supply chain operations with negative environmental impacts.

The dynamics of LMD requires an innovative bottom-up strategy to achieve an efficient and effective use of last mile delivery vehicles in a sustainable way. Importantly, integration of

167

last mile delivery into inner city development urban renewal projects requires appropriate urban planning inputs. In addition, last mile efficiency requires an integrated urban planning framework with an optimum logistics network, linking demand with supply. Such optimisation of transport network (e.g. full utilisation of vehicles; full load) will cut last mile delivery mileage whilst reducing environmental impacts. This is to consider the planning implication of last mile delivery to align with the full implantation of compact city model. Hence, efficiency, sustainability and cost-effectiveness are set out as the key goals for LMD.

7.2.2 Desirable Distribution Objectives (DDO)

Formulation of desirable distribution objectives is the second step in the setting up of the last mile delivery strategic framework. The outcomes under different scenarios from the workshop formulated six clear and well-defined desirable objectives that could help achieve the set goals. The six DDO’s are formulated from the four quadrants of the broader future scenarios (see Table 7.1). These DDO’s are from the Best-Worst, Worst-Worst and the Worst-Best scenarios (Figure 7.3). The Best-Best Scenario can be described as the ideal world where all things work well for last mile logistics and thus do not require any immediate intervention. Strategies backed up with actions are required to achieve the desirable distribution objectives across other scenarios.

Table 7.1: Desirable Distribution Objectives

Desirable Distribution Objectives DDO 1 Reduce gridlock/congestion associated with LMD DDO 2 Remove LMD delay DDO 3 Increase LMD productivity DDO 4 Increase daily LMD DDO 5 Reduce or towards a zero LMD environmental impact DDO 6 Increase LMD usage of transport network system

168

Best/Worst (Q1) Best/Best (Q2)

DDO 2 No immediate intervention DDO 4 required

DDO 6

DDO 1 SUPPLY DDO 2 DDO 1

INFRASTRUCTURE INFRASTRUCTURE DDO 3 DDO 3 DDO 4 DDO 5 DDO 5 DDO 6 Worst/Worst (Q3) Worst/Best (Q4)

LANDUSE INTENSITY Figure 7.3: Applicable potential outcomes and desirable distribution objectives for policy intervention in LMD

The thesis proposed five explicit strategies that would achieve the Desirable Distribution Objectives. The purpose of these strategies is to address the challenges identified for the future scenario resulting from decay in constraints within the Infrastructure Supply and continued Landuse Intensity through compactness and increased utilisation of land use. The strategies are expected to mitigate challenges related to gridlock and decrease in usage of transport network as well as increase delay. The strategies will also mitigate challenges related to productivity loss and reduction in daily delivery as well as worst negative environmental impact of last mile delivery.

7.3 Last Mile Delivery Strategies

The developed strategies combine different possible actions that could be considered by administrators, operators of last mile delivery and the community to help manage the provision of city logistics. The strategies are formulated to prevent the realisation of worst- worst LMD scenario and the concomitant challenges.

169

The localised spatial plan presented in Figure 7.4 illustrates some of these strategies with a particular focus on Maribyrnong City. The use of any or combination of the strategies should be guided by the spatial layout and investment available to supply logistics infrastructure to support LMD. The spatial plan identifies a potential location for establishing an Urban Consolidation Centre (UCC), delineated last mile delivery corridor and clearways to expedite movements of last mile delivery vehicles. The proposed location of the UCC is recommended given the proximity to Footscray Metro Activity Centre and adjacency to railway station and yard. The proximity will assist in the use of multi-modal transport in tackling future growth in regional freight volume into the Footscray Metropolitan Activity Centre. From the UCC, local freight to the Footscray Metropolitan Activity Centre and the Highpoint Activity Centre can be sustainably distributed out via smaller non-motorised modes like cargo bike or electric bike.

Figure 7.4: A localised spatial plan to augment the efficiency of last-mile delivery in Maribyrnong City

170

7.3.1 Landuse Zoning Strategy

Landuse zoning is a planning tool used by planning authorities to designate permitted uses of land and control development. The Victoria Planning Provision and the different local government planning schemes are the policy documents that contain these controls.

The purpose of Land-use Zoning Strategy (LUZS) proposed in this study is to regulate last mile delivery vehicle movements across Metropolitan Melbourne and integrate last mile delivery consideration into the landuse planning process. This will include delineation of a zone for logistics functions and integration of last mile delivery vehicle demand with off- street loading and unloading provision availability. Such land use zoning strategy requires demarcating last-mile delivery logistics zones to facilitate freight movements and operational requirements particularly along principal and major activity centres. Such should include reserving areas for logistics provision and for off-street loading and unloading operations into the Victorian Planning Provisions and relevant Planning Schemes.

Reintroduction and strict adherence to the loading and unloading provision (Clause 52.07 of the Victoria Planning Provision) and introduction of logistics zoning overlay through planning scheme amendment is a necessary step towards a fully integrated planning framework. Logistics zoning system can be developed to demarcate areas of high intensity of logistics use to improve freight flows and last-mile delivery to retail businesses within the CBD and higher-order activity centres.

As all policies are land referenced, the strategy will assist in solving problems related to gridlock and congestion remove last mile delivery delay and increase usage of transport system network system. Associated actions in achieving these objectives are graphically presented in Figure 7.5. The use of underground loading area, and delivery tunnel and planning scheme amendments are suggested actions that can be deployed to implement this strategy.

171

LMD Objectives Actions Reduce gridlock and Underground loading congestion associated with and unloading area last mile delivery Landuse zoning Remove last mile delivery Delivery tunnel delay strategy

Plannig Scheme Increase LMD usage of amendment transport network system

Figure 7.5: Landuse zoning strategy and associated actions

For example, the strategy can be applied within Maribyrnong, through the implementation of demarcation between the Footscray Metropolitan Activity Centre and the Highpoint Activity Centre. This can be tested and applied to other councils within Metropolitan Melbourne. In addition, designation of off-street loading/unloading and curb side loading zones including cut-outs (Figure 7.6) of wide sidewalks for delivery or waiting when loading area is still occupied is important in this strategy.

A new logistics zoning system can be introduced and legislated to demarcate logistics zones to differentiate the scale and intensity of freight movements.

Figure 7.6: Example of curb-side cut-out (Source: Giannopoulos, 2016)

172

7.3.2 Last Mile Delivery Corridor Strategy

Last Mile Delivery Corridor Strategy (LMDCS) relates to the provision of freight routes that goes through other constrained land uses to facilitate faster last mile delivery between proximate the CBD and activity centres or between activity centres.

Such last-mile corridor strategy can be implemented along the main arterial networks through linear freight routes to improve last-mile efficiency between key business hubs (Boyer et al., 2009; Srai et al., 2014). Use of such dedicated delivery corridor with time-window based loading/unloading will help reduce the environmental footprint of last-mile delivery, ease traffic bottlenecks and conflicts between last-mile delivery truck drivers and other road users. Designation of a last-mile corridor would also likely reduce last-mile delivery operational costs.

A parking clearway rule can be implemented in designated LMD routes to reduce the traffic constraint bottlenecks in peak hours. Traffic diversion technique to traffic flows can also be developed to divert freight traffic from more congested routes to low transportation network impedance routes. This strategy can be implemented to reduce transport network impedance to LMD within a compact city model.

This strategy will target four last mile delivery desirable distribution objectives (Figure 7.7). The use of underground loading and unloading area, use of delivery tunnel, provision of dedicated road corridor with increased speed limit and removal of road side parking to enhance last mile delivery vehicles are action proposed in this strategy.

For example, within Maribyrnong City, certain types of vehicle may be permitted in specific zones of the Footscray Metropolitan Activity Centre only at specified times, or they may simply be excluded from the area all together. These dedicated LMD routes especially along Hopkins, Barkly and Irving Streets can be made clear during the peak-hours via implementing a clearway policy where no parking permitted along the road side or curb.

Optimisation of last-mile delivery networks within the Footscray Metropolitan Activity Centre in inner city region can potentially solve problems caused by increased commercial vehicle movements (Fusco et al. 2003; Taniguchi et al. 2003). In addition, Off-Peak hour and night delivery can be included in this strategy. This should involve silent trucks to operate within the city centre in late hours to avoid road congestion and manage noise

173

pollution. Incentives for operators who agreed to this delivery can be taken into consideration to supplement cost incurred by operators in hiring staff to man the premises for the delivery.

Actions Underground loading LMD Objectives and unloading area Reduce gridlock and congestion associated with last mile delivery Delivery tunnel

Remove last mile delivery delay Last mile corridor Dedicated road corridor strategy

Increase daily LMD Increased Sped limit

Removal of road side Increase LMD usage of parking transport network system Incentives for LMD operators

Figure 7.7: Last mile corridor strategy and applicable actions

The Somerville Road and Rosamond Road in Maribyrnong City can be dedicated clearway area corridor to facilitate delivery between Footscray Metropolitan Activity Centre and the Highpoint Activity Centre.

7.3.3 Distribution Network Strategy

Distribution Network Strategy (DNS) in city logistics aims at optimising the distributed network of storage, warehousing and transportation systems as an integrated system (Toth and Vigo 2002; Lee and Jeong 2008; Crainic et al. 2009). Last mile delivery often experiences inefficiency due to unconsolidated deliveries, delays at loading bays, lower load factor and empty running. Urban Consolidation Centres are built as a distribution network strategy to improve city logistics provisions through freight consolidation, tasks coordination and resource sharing. The Novelog project by European Commission demonstrates the

174

application of this strategy. The UCC will enable goods to be de-bundled and redistributed for delivery into fewer fuel-efficient delivery vehicles.

This strategy integrates people, facilities and transportation infrastructure as a single unified logistics system. An Urban Consolidation Centre can be established in Maribyrnong to facilitate delivery of freight in Footscray and Highpoint. This can help improve supply chain coordination and a reduction in the number of trucks driving into or through the activity centres.

At the distribution centres, goods can be de-bundled, reorganised, stored and redistributed via a more sustainable form of transportation within the city centres and activity centres. The delivery from the distribution centres can be carried out through the use of electric vehicles and manual cargo bike. Where a permanent distribution centre is not appropriate or less desirable, a mobile depot can be established to distribute goods with greater agility to adapt to unpredictable shift in demand and changing traffic conditions over time and space. Footscray Market and the Little Saigon Market with Maribyrnong City can take advantage of this type of delivery. An efficient functioning of UDC, however, necessitates logistics collaboration between key stakeholders which should be based on openness, risk sharing and mutual trust as well as shared reward to help mitigate potential delay in delivery of goods in the last-mile component of the supply chain (Naesens et al. 2009; Lindawati et al. 2014; Park et al. 2016).

The aim of recommending this strategy is to reduce or achieve a zero LMD environmental impact and provide low cost delivery. The use of underground loading and unloading complemented with ITS will assist in achieving the objective and tackle the challenges associated with the problem (Figure 7.8).

175

Actions Underground loading and unloading area LMD Objectives Urban Consolidation Increase last mile delivery Centre productivity and reduce cost Distribution Network Strategy Shared economy Reduced or zero environmental impact Collaboration

Intelligent Transportation Systems

Figure 7.8: Actions to mitigate LMD challenges through Distribution Network Strategy

7.3.4 Multi-Modal Use Strategy

Multi-modal use strategy relates to the use of at least two or more modes of transport. This could be via road and train or trams. This strategy could include the use of trucks, train or trams, cargo bikes or other non-fossil fuelled vehicles for different part of LMD. Crawford, (2000) proposed an urban rail or tram system to be incorporated into the last mile delivery system. Deliveries can also be through tunnel into underground loading docks situated within Activity Centres or at the Urban Consolidation Centre from where deliveries can be made to shopping centres via cargo bikes, electric vehicles and human powered pushcarts. Establishing distribution centres near the rail route will allow the use of existing train services for delivery and intra-modal transfers between the distribution centres and the retailers – final destination (Toth and Vigo 2002; Lee and Jeong 2012; Crainic et al. 2009).

The implementation of this strategy will reduce gridlock and congestion arising from the increased number of inbound LMD vehicles into the CBD and remove LMD delay. More importantly, the strategy has the advantage of significantly reducing the total dependence of

LMD on road, which will ease congestion, enhance road safety and reduce CO2 emission from LMD vehicles (Figure 7.9).

This strategy can be achieved through freight consolidation, delivery through tunnel and use of Intelligent Transport Systems.

176

LMD Objectives Reduce gridlock and congestion associated with Actions last mile delivery Urban Consolidation Centre Remove last mile delivery delay Multi modal Transportation Delivery tunnel Increase last mile delivery productivity and reduce cost Intelligent Transportation Systems Reduced or zero environmental impact

Figure 7.9: Actions to achieve the objectives of the multimodal land transport strategy

Torino in Italy has utilised rail station located close to the urban CBD to organise and allow temporary warehouses (Pronello et al., 2017). Such opportunity allows for a transitional transfer arrangement through a multi modal last mile delivery system.

With increased expectation in last mile delivery requirements, traditional delivery measures will unlikely to satisfy the future requirements of goods. The use of drones and robots (Giannopoulos 2009; Hoffman and Prause 2018) would transform delivery of goods in the face of ‘internet of things’. Drones for delivery (Figure 7.10) will offer considerable benefits with ability to travel at a faster and sustainable speed without encountering the hurdles imposed by current built environment impedance.

The use of multimodal strategy including the use of drones as an alternative model requires further study, deeper understanding and feasibility assessment prior to the full implementation of the technology.

7.3.5 Stakeholder Engagement Strategy

The stakeholder engagement strategy involves administrators, operators and community in augmenting LMD decision-making process. Key stakeholders are capable of exerting positive influence and persuade the industry to ensure cost-efficient, and sustainable LMD through their actions. Particularly, this strategy places greater emphasis on the shift of interest of stakeholders with high power towards an improved LMD system. These are stakeholders in

177

the quadrant 1 of stage 5 in the research design. Stakeholders in the quadrant hold high power in making change through devising and deploying policies and regulation in the management of city logistics. Federal, State and Local Government and Transurban could shift in that direction to take greater interest in shaping city logistics provisions.

Consideration of this strategy to tackle the worst-worst scenario is necessary to remove LMD delay, increase usage of transport network by last mile delivery vehicles, reduce or remove environmental impact and assist in low cost and efficient LMD (Figure 7.10).

LMD Objectives Actions

Remove last mile delivery Encourage shared delay economy

Planning Scheme Reduced or zero amendment environmental impact Provide incentives for Stakeholder sustainable LMD Increase last mile delivery productivity and reduce cost Engagement Public and Private LMD partnership

Low cost and efficient LMD Set performance emission targets

Increase LMD usage of transport network system Shift of interest

Figure 7.10: Action to achieve the outcome objectives associated stakeholder Engagement in Decision Making

The change in the interest of federal, state and local governments and Transurban from quadrant 1 (high power/low interest to quadrant 2 (high power/high interest), to join Vicroads and Lobbyist is necessary to address the worst-worst last mile delivery future scenario (Figure 7.11).

Required actions from these stakeholders relates to encouraging shared economy and provision of incentives for operators and retailers who are willing to embrace sustainable delivery or late night delivery.

178

In addition, the identification of a “freight person/unit” as a focal point in key agencies like Vicroads and Local Government will be a way to facilitate last mile delivery discussion and generate appropriate stakeholder engagement and partnerships. Also, educating elected official and creation of awareness of the last mile logistics importance to the overall economy and liveability of the city will likely change their perception on last mile delivery. Such stakeholder engagements have been used in New York (Capital District Transportation Committee, Albany, New York), and London (The Central London Freight Quality Partnership - CLFQP), in facilitating the implementation of freight initiatives.

Q1 Q2 Federal Government High Power High Power Transurban Public State and Transport Local Victoria Government Vicroads Low Interest High Interest Port Authority Lobbyist Q3 Q4

POWER Low Power Low Power

Low Interest High Interest

Low High INTEREST Figure 7.11: Shift of interest of federal, state and local government and Transurban to cause change in last mile delivery

7.4 Possible Actions (PA)

Policy intervention or action is a combination of schemes designed to achieve an objective and produce an improved change or outcome. Achieving the key desirable distribution objectives requires combination policy actions. Given the nature of last mile delivery, one policy measure intervention cannot tackle different objectives of the challenges. Hence a one stop method is not possible in solving the current last mile delivery problems. Strategies to

179

implement these interventions to achieve the objectives would include different policy interventions highlighted in section 7.2.1.

The following possible policy actions (PA) are proposed for inclusion in any combination of strategies in achieving the key distribution objectives.

7.4.1 PA 1: Regulating the Size of Vehicles and Delivery Time

This policy intervention measure aims to regulate the size of vehicles and delivery time. It relates to the control of specific times when vehicles can make delivery as well as loading factors, and fuel emission elements (Daniel et al., 2013). This measure also seeks to regulate and restrict delivery of certain vehicles only at specific times or total exclusion of some vehicles from access to CBD and activity centres within the city. The restriction will assist in mitigating gridlock and congestion and decrease in usage of transport infrastructure. In other cases, shifting of peak hour delivery to non-peak hours has been implemented in some cities in Europe to curb traffic congestion by last-mile delivery vehicles. Operators who adapt to these non-peak deliveries are provided with incentives through Pierpass OffPeak programme. In addition, the use of fuel efficient small size vehicles or electricity powered vehicles will mitigate negative environmental impact of high power vehicle delivery trucks.

This possible action has been noted not to completely solve last mile delivery problem (Eric et al., 2007), given the need for various demands of goods at varying time of the day. Furthermore, additional cost of attending to night delivery will put extra burden on retailers. Also, environmental impact through Co2 contribution is not likely to be mitigated by this method because there are no day or night difference in Co2 emission from last mile delivery trucks. The federal government through the state and local government policies and regulations can use their regulatory power to stimulate actions and change in association with other complementary strategies.

In addition, Vicroads can assist in the implementation of this method. However, drivers and business owners with high interest and low power are more likely to be negatively affected by the use of this method. This might also increase daily delivery lead time and reduce gridlock resulting from last mile delivery vehicles. Encouraging sustainable vehicle by providing incentives or subsidies to use them by government at different levels will go a long

180

way in inspiring last mile delivery operators to use this type of vehicles. But it will come with the challenges associated with subsidising an industry from publically funded initiatives.

7.4.2 PA 2: Consolidation of load and Load Factor Efficiency

Consolidation of load and load factor efficiency is another Possible Action that can form part of the strategy to tackle LMD challenges and to achieve the Desirable Distribution Objectives (DDO) (Danielis et al., 2013; Reisman and Chase, 2011). Load factor refers to the total percentage of goods that last mile delivery is filled with at any particular time of travel. Load consolidation can be implemented for delivering goods to premises that require multiple shipments a day from various suppliers (Seattle Urban Mobility Plan, 3). This is then possibly helps reduce the declining load factor issues, the number of delivery vehicles on the road, reduce the amount of delivery travels required, increase usage of transport network infrastructure and landuse in a worst-worst scenario.

The shortfall of this possible action is that delivery vehicles may require to be specially designed and compartmentalised to allow load consolidation that requires different loading techniques of temperature (Danielis et al., 2013). Also, the location of point to consolidate the load for increased load factor may not be close to achieve efficient and cost effective pickup and drop off at consolidation and delivery locations.

Truck association with high power, Business Owners and Drivers are in position to implement this possible action. Consolidation of deliveries through increased load factor may lower cost and travel repetitiveness.

7.4.3 PA 3: Urban Consolidation Centre (UCC)

The basic drive for the adoption of an UCC strategy is to limit part load delivery to city centres or apartments therein. It aims at providing facilities to allow appropriate consolidation of last mile delivery vehicle with a high level of load utilisation. The basic characteristics in the description of UCC are its location and geographical proximity the area such UCC serves. This geographic area served can be a whole city, local or regional shopping centres or a principal activity centre where consolidated deliveries need to be carried out for pick up or delivery (Browne et al., 2005). The advantage of UCC is in vehicle trips, distance coverage, fuel consumption and subsequent emissions from the trips, distance and fuel consumptions.

181

Such an UCC model or its variants are implemented in Europe (especially in the United Kingdom, Netherlands and Germany), Canada, USA and Japan. These commercially operated UCC in the UK and Netherlands or site specifics in United Kingdom and the cooperatively operated ones in Germany and Japan (Bjorklund and Johansson, 2018). The commercially operated UCC requires payment for delivery/pickup of goods, site specifics are associated with those under a single entity or organisation like an airport authority or a large commercial building; In contrast, the cooperatively owned UCC represents carriers coming together to consolidate freight and distribute cooperatively. Distribution can be carried out on cargo bicycles or smaller-sized low-load capacity electric vehicles. In addition, the UCC can serve as a bundling/de-bundling centre and for storage purpose due to its proximity to the CBD or shopping centres. The use of such UCC can include a mobile depot (Figure 7.12) located near the CBD or activity centres.

Figure 7.12: Mobile UCC (Source: Straightsol project 2015)

The SMILE project in Barcelona city in Europe has piloted the use of UCC method where operators deliver goods to a local UCC called transhipment point. The Smile project emphases state-of-the-art solutions to improve the efficiency of last mile delivery through reducing greenhouse emission associated with last mile delivery.

From the transhipments point, goods are delivered to the owner through cargo bike or electric electric tricycles in the CBD, a form of micro last mile delivery management (Figure 7.13).

182

Figure 7.13: Smile Project typology in Barcelona (Source http://piraeus.gov.gr/wp- content/uploads/2018/11/1st-newsletter-of-the-SMILE-Project..pdf)

The overall process and workability of this use of UCC can be visualised as presented in Figure 7. 14. Goods from different trucks are sorted and consolidated within the UCC. Deliveries are ‘Truckload IN’ and out in ‘Consolidation OUT’ for delivery.

Figure 7.14: Typical operation of Consolidation (Source: http://www.bestfact.net/wp- content/uploads/2016/02/2-143_BESTFACT_CL2_QuickInfo_SMILE.pdf)

Despite the laudability of this Possible Action, Schoemaker (2002) identified the factors of failure of the Laiden UCC in Germany to include the following:

 the location of such UCC, too far away from the highway or from the CBD where smaller sustainable means can be used for delivery;  lack of supporting policy interventions;  the unwillingness of last mile delivery operators to use such and proliferation of UCC;  problem of speed of electric vehicles or cargo bicycle slowing down traffic; and  financial feasibility of such UCC due to lack of volume.

In Bristol, only 20% of the shop owners participated in the use of UCC. This is because of the lack of strict enforcement by the city municipalities (Patier and Browne 2010). The use of this PA in an urbanised environment will require avoiding the failure factors of potential failure highlighted by Schoemaker, (2002) with consideration of locational factors backed with favourable policies and regulations. To incorporate this method, the federal, state and local government, transurban in partnership with truck association, business owners, port

183

authority and local communities including developers of high density housing within the designated activity centres can co-contribute to building of such UCC for pick-ups and deliveries.

7.4.4 PA 4: Shared Economy Business and Information Technology Systems

Shared economy focuses on utilising underutilised assets. The shared economy model has been successful in car sharing and can be applied to last mile delivery. The dominance of third party operator in last mile delivery is not new, therefore utilising underutilised assets in the form of collaboration of citizenry through crowd sourcing is a method that can be implemented in the context of last mile logistics. This model would allow individuals to deliver a package on their daily commute. The bringBuddy in Germany and the You2You French models are examples of the shared economy model. Shared economy business model can be incorporated into collaboration in last mile delivery operators and subsequently reduce carbon footprint of last mile delivery. The proposal for this model is what Banker, (2016) described as ‘Ubernization of last mile’.

Truck association, local community, business owners and other road users are in position to be involved in such shared economy, which intends to reduce the number of last mile delivery vehicles, especially where load factor is low. This method has to be in conjunction with appropriate information technology (Figure 7.16).

Opportunity for a real-time monitoring via Global Positioning System (GPS) and Intelligent Transportation Systems (ITS) is a tool that last mile delivery operators can utilise to track the movement of freight. Such real-time assessment can be used to locate availability and location of real time loading area within or in proximity to shopping centres. This is especially in consideration of the future of Internet of Things (IoT) and application of big data. Information technology will provide opportunities for monitoring of distribution of last mile vehicles and online communication for routing and scheduling changes that can occur between delivery vehicles (Giannopoulos, 2009).

Local government has been using real time ITS tools to determine the availability of car parking spaces in cities and this can be developed to finding the availability of loading and unloading zones. Drivers and business owners are in a position to benefit from the use of such GPS and ITS tools. A reduction in number of vehicles driving through the city when

184

required unloading space is occupied will be eliminated through real time relay of information and knowledge of loading space availability.

The use of crowd sourcing in B2B delivery is constrained by weight and requirements of heavy lifting equipment. In addition, local councils may not be able to commit funding to implement this method. Business owners and last mile delivery operators are the one who can finance the infrastructure required to build the IT system.

Figure 7.15: Information Technology and Crowd sourcing in last mile delivery (Source https://kevingue.wordpress.com/2014/06/17/the-social-implications-of-crowdsourced- delivery/)

With increased expectation in last mile delivery requirements, traditional delivery measures will not likely satisfy the future requirements of goods ordered online due to the volume. The use of drones and robots (Giannopoulos 2009; Hoffman and Prause 2018) for freight delivery in the face of ‘internet of things’ might be the ultimate transport system for the future of last mile delivery. Drones for freight delivery (Figure 7.17) will offer considerable benefits with the ability to travel at a faster and sustainable speed without encountering the current built environment impedance. However, the use of drones requires future research, innovation and flexible regulatory regime to support the wider adoption of technology in a complex urban environment.

185

Figure 7.16: Use of drones for delivery (Source: https://www.mckinsey.com/industries/travel- transport-and-logistics/our-insights/how-customer-demands-are-reshaping-last-mile- delivery)

7.4.5 PA 5: Underground Loading and Unloading Area

Designing an underground loading dock is now becoming part of urban design and planning of large city centres. Delivery through underground loading areas is an advanced solution for last mile delivery problems (Koshi et al., 1992). This has been estimated to reduce negative environmental impact of last mile delivery on the built environment (Koshi et al., 1992; Oishi, 1996; Visser 1997; Taniguchi and Heijden, 2000).

This PA has been employed for delivery for Emporium within Melbourne CBD. This underground delivery facility serviced 225 stores receiving approximately 100 deliveries per day. The facility is equipped with two truck lifts which can accommodate six trucks at a time. The facility operates through an online booking system allowing drivers to 45 minutes time for unloading. The online booking allows last mile delivery operators to know exactly when to arrive as well as guarantees access on arrival.

The federal, state and local governments can implement this through legislation. Business owners, Truck association and VicRoads can encourage such underground loading docks from the conception and commencement of projects. Developer and Business Owner can incorporate such delivery technique into the design of the development project.

186

7.4.6 PA 6: Delivery through Tunnel

One other possible action is the use of delivery tunnels. This measure in last mile delivery is emerging yet limited to few cities. Last mile delivery is often carried out through underground pipes with ‘Elon Musk-style electromagnet technology’ (Tingle, 2018). This technology would allow for shell cases inside the hyperloop to go faster and move with little or no friction. The aim of underground delivery tunnels is to use automated vehicles to travel underground for deliveries in the city. The Chicago Tunnel Company and the Post Office Railway uses underground rail for over 26 miles for delivery. Chicago Tunnel Company and the Post Office Railway have since closed. This is because the cost of running the associated equipment far exceeds movement of goods over land.

In recent times, similar underground project to include last mile freight is proposed in Helsinki and Switzerland. The Cargo Sous Terrain project in Switzerland is proposing an underground tunnel system. The tunnel will be 66 kilometre long, connecting 3 cities / logistics centres in Switzerland (Soluthurn, Bern, and Zurich). In line with this, shopping centres have been incorporating underground loading and unloading areas for last mile delivery. A typical tunnel movement of freight is presented in Figures 7.18 and 7.19.

Implementing this Possible Actions (PA) will require huge funding from stakeholders with high power given the huge financial implication. The operators of toll roads can invest in this type of delivery through a public-private partnership funding arrangement.

Figure 7.17: Typical use of tunnel for delivery to city centre (Source: https://www.cargoforwarder.eu/2018/10/29/jd-com-to-study-underground-urban- parcel-delivery-network/)

187

Figure 7.18: Typical use of tunnel for delivery to city centre (Sources: http://www.logisticsmatter.com/2017/05/11/freight-going-underground/)

7.5 Summary

This chapter developed a strategic framework for an efficient, sustainable and cost-effective last mile logistics. It proposed five strategies that can be implemented to mitigate LMD challenges. Six desirable distribution objectives were developed with actions suggested to achieve the objectives. To demonstrate these strategies and possible actions, this research developed a spatial plan of Maribyrnong City as a case study to mitigate the less desirable outcomes of plausible LMD scenarios. However, no single strategy can be applied exclusively to the entire Metropolitan Melbourne.

Possible Action to be included in the strategies includes regulating the size of vehicles and delivery time; consolidation of load and load factor efficiency; use of Urban Consolidation Centre; Shared Economy Business and Information Technology Systems; Underground Loading and Unloading Area; and delivery through tunnel. The implementation of these strategies is dependent on the economic viability, urban design and the regulatory environment which will inform the feasibility of specific strategy or combination of strategies.

The final Chapter 8 presents the conclusion with discussion on limitations and how this research contributes to knowledge.

188

CHAPTER EIGHT

CONCLUSION

8.1 Introduction

The ‘last mile’” logistics in cities is becoming increasingly complex. As global sourcing and online shopping expand, last-mile logistics of both inbound and outbound commodity chains within cities will become increasingly challenging. To tackle these challenges, this study developed strategies to mitigate the risk of last-mile delays to reduce transportation network impedance. This study has estimated and mapped the potential last mile delivery impedance and constructed plausible future last mile delivery scenarios. This chapter presents a summary of the key findings from chapter 1 through to 7. The chapter also contains the major contributions to the body of knowledge and potential implications of the core findings for industry and government as well as major limitations and future direction.

This chapter is organised into seven sections. A brief introduction is provided in Section 8.1 and Section 8.2 presents the key findings of this study. Section 8.3 revisits the research questions developed in Chapter 1. The contributions and implications of the study are presented in Section 8.4. Section 8.5 highlights the limitations of this research, while the direction for future research is detailed in Section 8.6. Section 8.7 concludes this study by providing the final concluding remark on the thesis.

8.2 Key Findings

This study has implemented a five-stage research design that seamlessly integrated Scenario Thinking and Theory of Constraints to formulate plausible LMD scenarios. In addition, GIS techniques were applied to map last mile delivery impedance at a local area level. This section highlights the key findings from the analyses, which have formed the basis for the development of strategic framework to address the projected outcomes for LMD scenarios.

Overall, the key findings of the research are collapsed and presented under four themes, which directly answer the research questions set out in chapter 1.

189

8.2.1 Six Dimensions collectively represent the Last Mile Delivery Constraints

Thirty-four key constraints were identified by stakeholders, and were clustered into six dimensions; namely Freight Infrastructure, Landuse Intensity, Infrastructure Supply, Infrastructure Sharing, Intersection Control and Human Behaviour.

The constraints within these broader dimensions of city logistics are found to be among those which are holding the last mile delivery from being efficient, sustainable and cost-effective. The constraints positively or negatively impede LMD. Higher intersections, traffic lights per road segment; lesser lanes, speed limit, road width; availability of bus lanes, bicycle lanes, tram routes, boom gates; and non-availability of loading area exerts negative effect on LMD. On the other hand, lesser intersection, traffic lights; higher lanes, speed limit, road width; non-availability of bicycle lanes, bus lanes, tram routes; boomgates; and availability of loading exerts positive effect on LMD.

8.2.2 Last Mile Delivery Impedance varies across Metropolitan Melbourne

LMD impedance varies spatially across the Metropolitan Melbourne. Population density and proximity to activity centre act as major drivers of LMD impedance. Areas with higher population density particularly residential zones and within a close proximity to activity centres are found to be areas of high level of impedance. LMD impedance levels decline outward from Metropolitan Melbourne inner towards middle and outer rings. Levels of higher impedance are prevalent within the inner ring with high population density. More medium to low impedance are recorded within the middle ring and the outer ring with low population density. Correlation coefficient of LMD impedance level and population density reveal a strong positive relationship.

There is a link between the LMD impedance and the attributes of compact city model which reflects in the form of increased population density and activity centres. These areas within closer proximity to activity centres are found to hold high LMD impedance level. These high impedance areas are more likely to be congested as a result of shared infrastructure with tram lanes, railway boomgate delay, routes connecting principal and major activity centres.

190

8.2.3 Future of Last Mile Delivery is shaped by Infrastructure Supply and Landuse Intensity

Infrastructure Supply and Landuse Intensity were identified as the key drivers that would shape the future of last mile delivery. These two dimensions were derived through Scenario Thinking process and utilised in the generation of four plausible LMD scenarios. These two dimensions are of the highest impact (impedance) with highest uncertainty (low certainty). The future of last mile delivery would likely to be dependent on the best and worst interplay between these two dimensions.

The best-best scenario of the two dimensions would result in an increase in road usage by last mile delivery vehicles, coupled with decreased congestion, increased productivity, lowering on delivery cost and lowest environmental impact. Planners and Logisticians need to consider this scenario as ‘what to change to’.

On the other hand, the worst-worst scenario will result in gridlock and decrease in usage of transport network, increase delay in last mile delivery, productivity loss and increase negative environmental impact. This worst future is what stakeholders and operators in last mile delivery will want to avoid. The worst-worst scenario necessitates the need for immediate intervention and development of a strategic plan to mitigate the incidence of damaged products, congestion, last mile delivery stagnation and ageing infrastructure.

8.2.4 Stakeholders with Power and Interest can cause Positive Change in Last Mile Delivery

A total of 14 stakeholders were identified who have direct interest or power in the field of city logistics. These include Federal, State and Local Government, Transurban, Truck Association, Public Transport Victoria, VicRoads, Port Authority, Lobbyist Group, Other Road Users, Business Owners, Local Community, Drivers and Non-Governmental Organisations. Federal, State and Local governments and Transurban are of high power that can cause positive change in last mile delivery and ensure efficient, sustainable and cost effective last mile delivery.

191

Also, VicRoads and Traders Association represent the high power and high interest to help lobby and campaign for more infrastructure investment and technological innovation to favour last mile delivery in emerging city compactness.

8.3 Addressing the Research Questions

The overall aim of this research is to construct future last mile delivery scenarios, examine and map the potential last mile delivery impedance within Metropolitan Melbourne transport network using Scenario Thinking methods based on the constraints of urban planning and transport network.

The discussion in the following section presents how the research questions were answered as discussed in chapters 3, 5, 6 and 7.

 What are the key built and regulatory constraints of last mile delivery impedance in an urban setting?

This research question is addressed by the constraints identified through interactions of the built environment, planning and transport systems and last mile delivery via ST workshop. 34 key constraints that impede last mile delivery in an urban setting were identified which are discussed in chapters 3 and 5. The constraints limit last mile delivery from achieving its goal. These constraints apply to any urban setting experiencing growing urban compactness through increased density and land use mix.

 What are the key dimensions that drive the future of last mile delivery scenarios?

The 34 constraints that impede last mile delivery identified in research question 1 were further clustered into six high-order dimensions through consensus building process and were presented in chapter 6. The six dimensions collectively capture all the identified constraints. These include Freight Infrastructure, Landuse Intensity, Infrastructure Supply, Intersection Controls, Infrastructure Sharing, and Human Behaviour. Specifically, Infrastructure Supply and Landuse Intensity are the two key dimensions that will drive the future of last mile delivery scenarios. RQ2 is therefore answered in stage 2 of the research design.

192

 What are the likely outcomes of the projected future last mile delivery scenarios?

This research question is answered by formulating the plausible LMD scenarios through the combination of the best and worst interactions of Infrastructure Supply and Landuse Intensity dimensions. The broader outcomes of each of the four scenarios were derived from two key dimensions identified in research question 2. This RQ is addressed in Chapter 6 of the thesis, which outlined likely outcomes of the projected future LMD scenarios.

 What strategy can be formulated to mitigate potential transport network impedance to last-mile delivery?

Identifying problems or constraints will be futile without efforts to mitigate their level of impedance on LMD. This necessitates the need for ‘how to cause change’. This research question is answered in chapter 7 of the thesis. Five key spatially explicit logistic strategies are recommended to address the challenges of last-mile logistics. These strategies are combination of various actions. The recommended strategies are logistics zoning strategy, last-mile corridor strategy, adaptive distribution network strategy, multimodal strategy and stakeholder engagement in decision making. The development of these strategies directly addresses this RQ.

8.4 Contribution of the Study

This thesis has made significant theoretical and methodological contributions to the current body of knowledge. This study shifts the traditional focus from a heavy reliance on guesswork or quantitative models to incorporating views and perceptions of city logistics stakeholders in portraying how the future of last mile delivery will eventuate if things remain the same. The key contributions and implication of the study are presented in the following sub-sections.

8.4.1 Theoretical Contribution

The approach adopted in this study and the results produced have improved the theoretical understanding of the complex interactions between planning and transport systems and LMD within a compact city-driven model. The conceptualisation of LMD impedance as a measurable phenomenon is a major theoretical contribution of this study. The concept of LMD impedance is theorised by using the Theory of Constraints that underpin the scale and

193

intensity of LMD impedance. Traditionally, the logistics and supply chain management discipline tend to largely focus on operational aspects of city logistics; whilst urban planning concentrates on planning and regulatory functions. Incorporation of both operational and planning/regulatory constraints to estimate levels of LMD impedance is novel. The theoretical approach adopted in this study is innovative as it has seamlessly integrated two discreet theories; namely the Theory of Constraints and Scenario Thinking, in order to formulate plausible and possible LMD scenarios.

This study extended the application of the principles of TOC to establish planning and transport networks constraints of LMD. It has implemented the thinking process framework relating to ‘what to change’, ‘what to change to’ and ‘how to cause change’ with ‘who to cause change’ in the context of last mile logistics.

8.4.2 Methodological Contribution

This is the first study that estimated and mapped LMD impedance as a geographical phenomenon. The locational attributes included in the study represents notable spatial disparity within a geographic space. An integration of GIS functions into the Scenario Thinking approach is also methodologically innovative. Visual capability has not only added value to the ability of Scenario Thinking approach to create likely scenarios but also the capability to illustrate spatially-varying future outcomes. Identification of key constraints will help in determining the overall assessment of B2B LMD literature from an urban planning perspective.

Efficiency and responsiveness of a delivery system are space dependent. Space affects the way customers and supply are distributed across the market, which in turn affects delivery window length. In this context, the spatial approach has provided a robust visual analytic framework to help structure logistics space in a way to enable the LMD network to optimally plan logistics services across different parts of the metropolis, in accordance to the demand and the regulatory environment. It was important to examine the impact of spatial context, as it is, as Harvey (1990) argues, a function of production and shows how space creates economic resistance to the effective and optimal movement of goods as argued by Hesse and Rodrigue (2004) and Keith and Pile (2013).

194

The LMD network impedance is highly dependent on the way their relative positioning are spatially organised and functionally integrated with urban infrastructure and planning restrictions. The spatial perspective of a LMD network is critical to the delivery of products as it regulates the way elements of a logistics network are arranged, structured and organised which, in turn, affects costs and delivery time (Chhetri et al., 2017). Constraints in the delivery time window impacts significantly on how the delivery trips are routed and scheduled.

8.4.3 Planning Implications

The built and regulatory environment contributes to mitigating the severity of LMD problem in urban areas. The use of land use controls as instruments to increase city compactness in strategic nodes/hubs is more likely to deter the movement of urban freight. This thesis has generated new evidence to expand the existing knowledge on the interplay of built environment, planning controls and last mile logistics, which would assist planners at broader urban planning level.

Overall, the finding from the research has the following planning implications:

 The mapped outputs reflect the spatial variability across Metropolitan Melbourne, which will aid decision-making in mitigating the potential delay in last-mile delivery through devising localised strategies such as dedicated freight corridors or time-bound deliveries in congested areas of road network.  The study has developed a spatial plan to enhance last mile logistics efficiency. LMD impedance mapping at a local area level provides a framework to develop strategies to mitigate the risk of service failure in areas of high impedance through timely planning interventions.  An improved understanding of plausible LMD future scenarios would help transport planners and logisticians to plan future investment strategies and operational plans to deal with the challenges of the worst/worst scenarios. Stakeholders with high interest and high power could play a critical role in mobilising support to lobby for improving the efficiency of LMD.  The down-scale mapping of LMD impedance at the street level adds an innovative urban planning dimension to decision making process in managing the complexity of

195

the growing importance of city logistics. Identification of difference in the levels of impedance will point to areas that require careful interventions to mitigate LMD bottlenecks.  A new logistics zoning system can be designed at a micro-level and legislated at a metropolitan level to demarcate logistics zones to improve freight movements. Such will be an innovative urban planning measure to mitigating LMD problems and add to the decision support tools to regulate city logistics.

8.5 Limitations

Despite the success of the research in estimating and mapping the potential LMD impedance, the limitations of the research should be acknowledged. These are highlighted in this section.

The first limitation relates to the Scenario Thinking approach, which has been criticised for inaccuracy in predicting the future or the occurrence of unknowables. The inability to forecast outcomes of the unknowables and the occurrence of such predictable behaviour and patterns limits the application of and usability of Scenario Thinking in business. Businesses need to estimate future demand to strategically plan production and predict for order fulfilment and any potential delays in the presence of service produce and demand uncertainty.

The second limitation relates to the process of building scenarios, which is driven by the configuration of two-dimensional space. The use of Infrastructure Supply and Landuse Intensity out of the six dimensions limits the robustness of the possible and plausible future of LMD. Incorporating more than two dimensions in scenario building would increase the complexity of building the scenarios and would be difficult to comprehend visually and cognitively.

The third limitation pertains to the dearth in LMD data, which has restricted the level of analysis presented in this study. The protection of data, cost of data collection and storage, data sharing, and the predominance of privately-owned data have all contributed to the availability and accessibility of data. Integration of origin to destination freight data, for instance, would enable estimating the actual impedance to LMD. These origins to destinations freight data are not readily available for researchers, specifically the movements of freight vehicles. The use of static measures of transport and planning constraints has

196

reduced the robustness of the transportation network impedance in confronting the challenges associated with city logistics. An integration of dynamic and real-time B2B data into LMD impedance would produce better results and more realistic outcomes for businesses and city planners.

Finally, the spatial approach adopted in this study is valuable for generating evidence for broader urban planning at metropolitan level, but its use is somewhat limited to assisting in daily logistics operational planning and scenario-based decision making. This is because of the dearth of data identified initially.

8.6 Future Research Direction

The agenda for future research will be specifically driven by the limitations of this study, which are listed in the previous sub-section.

Firstly, the scope of the Scenario Thinking workshop is restricted to a two-dimensional representation of the future of last mile logistics which may perhaps limit the impact of the interplay of multitude dimensions in formulating multiple scenarios and representations. In addition, the future last mile scenarios derived in the workshop are collectively created. However, there might be different perceptions of reality based on different stakeholders (e.g. drivers versus transport planners), which at times can be contradictory and too diverse to be converged through consensus. The future research will extend the scope of this work by exploring the possibility of building stakeholder-specific perception of the future of last mile logistics.

Secondly, the future research will aim to incorporate fleet movements based on GPS logged data to extract key constraints of last mile delivery. This will help capturing the dynamic and uncertain urban environment within which the last mile logistics operates. Incorporation of GPS logged data will need to be coupled with integrated technologies such as cloud-based solution combined with mobile applications and internet of things to connect transportation and warehouse management systems with vehicle sensors and other integrated devices for modelling.

Thirdly, future efforts will be directed to implementing the strategies developed to address the challenges identified in the future for last mile logistics. The effectiveness of different

197

types of LMD modes (i.e. Collection and Delivery Points, Attended Home Delivery and Reception Box) in different urban contexts will also be evaluated. Innovative strategies to optimise LMD will be developed and tested to assess the LMD performance across different urban designs and planning options. LMD options can be formulated to adjust changing consumer demand and behaviour, which often requires a trade-off between flexibility, security, speed and cost of delivery.

Fourthly, business to customer (B2C) and customer to customer (C2C) last mile logistics is likely to grow rapidly in the future. Businesses are encouraging customers to pick up a delivery from specialised boxes placed at different locations at a convenient time. Customers are also using crowdsourcing methods (e.g. BringBudy, You2You) to deliver goods for other customers. Future research will explore the scale and intensity of B2C and C2C LM logistics by characterising these transactions and the way these movements should be regulated to enhance business performance and reduce last mile distances and carbon-footprint.

Finally, the inclusion of expert choice modelling to evaluate different weights assigned to constraints would enhance the capacity and capability to visualise the last mile delivery impedance. Weight can also be generated through expert judgement or collected through survey of logistics managers, transport planners and drivers. Application of Analytic Hierarchy Process (AHP) with such weights to model last mile delivery constraints will be the next level of analysis. Application of AHP will allow a more rigorous statistical analysis of results and comparison of mapped constraints with or without weights.

8.7 Final Concluding Remark

The thesis developed a measurement framework to estimate and map the last mile delivery impedance and applied Scenario Thinking to build the future last mile delivery scenarios using the built and regulatory constraints of the urban environment. The research utilised a five-stage Scenario Thinking method that seamlessly integrated Theory of Constraints with Geographic Information Systems (GIS) to understand and analyse the provision of last mile delivery and the associated ‘critical uncertainties’. GIS was used to compute and visualise the potential transportation network impedance to last mile delivery within the Metropolitan Melbourne. Spatial data representing the key constraints of transportation network and

198

planning controls were used to compute the potential transportations network impedance to last mile delivery.

The results revealed that last-mile logistics in the future is possibly dependent upon which scenario eventuates out of the four formulated scenarios from the best-best through to the worst-worst scenario. The possible outcomes in the future are largely driven by Infrastructure Supply and Landuse Intensity. Stakeholders with power and interest could cause desirable changes through effective policy interventions and regulations with a cautious approach considering the growing complexity of city logistics in a large urban setting.

Localised strategies formulated to tackle the growing last mile problems would assist in the development of operational plans, deployment of geo-targeted future investment in logistics infrastructure and mitigation of the possible outcomes under the worst-worst LMD scenario. The study improved the future understanding of the complex interactions between transportation infrastructure and urban planning controls and last-mile logistics. The study provides evidence to inform urban planning process and developed strategies to enhance the efficiency of last-mile logistics, driven by low-volume, high-frequency and short-haul distribution of goods to end consumers in the context of large cities.

199

REFERENCES

Agunbiade, M. E., Rajabifard, A., & Bennett, R. (2012). The dynamics of city growth and the impact on urban land policies in developing countries. International Journal of Urban Sustainable Development, 4(2), 146-165. doi:10.1080/19463138.2012.694818 Agunbiade, M. E., Rajabifard, A., & Bennett, R. (2014). Land administration for housing production: An approach for assessment. Land Use Policy, 38, 366-377.

Ackermann, F. & Eden, C. (2011). Strategic Management of Stakeholders: Theory and Practice, Long Range Planning, 44(3):179-196

Aized, T., & Srai, J. S. (2014). Hierarchical modelling of Last Mile logistic distribution system. The International Journal of Advanced Manufacturing Technology, 70(5-8), 1053-1061.

Alexander, C. (1979). The timeless way of building (Vol. 1): New York: Oxford University Press.

Alho, A., Bhavathrathan, B. K., Stinson, M., Gopalakrishnan, R., Le, D.-T., & Ben-Akiva, M. (2017). A multi-scale agent-based modelling framework for urban freight distribution. Transportation Research Procedia, 27, 188-196. doi:https://doi.org/10.1016/j.trpro.2017.12.138

Alho, A. R., & e Silva, J. d. A. (2015). Utilizing urban form characteristics in urban logistics analysis: a case study in Lisbon, Portugal. Journal of Transport Geography, 42, 57-71.

Allen, J. and Browne, M. (2010) ―Sustainability Strategies for City Logistics.‖ In Green Logistics: Improving the Environmental Sustainability of Logistics, edited by Alan McKinnon, Sharon Cullinane Michael Browne, and Anthony Whiteing, 282-305. London: Kogan Page Limited.

Allen, J., Browne, M., & Cherrett, T. (2012). Investigating relationships between road freight transport, facility location, logistics management and urban form. Journal of Transport Geography, 24, 45-57. doi:https://doi.org/10.1016/j.jtrangeo.2012.06.010

Allen, J., Piecyk, M., Piotrowska, M., McLeod, F., Cherrett, T., Ghali, K. and Austwick, M. (2018). Understanding the impact of e-commerce on last-mile light goods vehicle activity in urban areas: The case of London. Transportation Research Part D: Transport and Environment, 61, 325-338. doi:https://doi.org/10.1016/j.trd.2017.07.020

Anastas, J. W. (1999). Research design for social work and the human services: Columbia University Press.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modelling in practice: A review and recommended two-step approach. Psychological bulletin, 103(3), 411.

200

Anderson, S., Allen, J., & Browne, M. (2005). Urban logistics––how can it meet policy makers’ sustainability objectives? Journal of Transport Geography, 13(1), 71-81.

Anderson, S. C. (2000). The globally competitive firm: Functional integration, value chain logistics, global marketing, and business college strategic support. Competitiveness Review: An International Business Journal, 10(2), 33-45.

Anderson, W. P., Kanaroglou, P. S., & Miller, E. J. (1996). Urban form, energy and the environment: a review of issues, evidence and policy. Urban studies, 33(1), 7-35.

Andre, D., Ginsburgh, V. and Thisse, J. (1987) On Existence of Location Equilibria in the 3- Firm Hotelling." The Journal of Industrial Economics 36.2, 245-252.

Andrews, F. M. (1989). The evolution of a movement. Journal of Public Policy, 9(4), 401- 405. Retrieved from http://www.jstor.org/stable/4007396

Angel, S., Parent, J., Civco, D. L., & Blei, A. M. (2011). Making room for a planet of cities.

ANRAM. (2014). ANRAM Attributes & Risk Factors: ANRAM Fact Sheet 2.

Australian Bureau of Statistics (ABS) (2018). Road Freight Movement, Australia. http://www.abs.gov.au/ausstats/[email protected]/0/7480A96EA1DF4527CA257EEC000EEB5 D?Opendocument

Australian Housing and Urban Research Institute (2015) Processes for developing affordable and sustainable medium-density housing models for greyfield precincts, Appendix 1, ISBN: 978-1-922075-76-5 precincts, Appendix 1, ISBN: 978-1-922075-76-5

Australian Road Rules Under the Road Traffic Act 1961. National Road Transport Commission. Version 8.12.2016, ISBN 0 7240 8874 1.

Azad, N., & Rostamnya, Y. (2012). Ranking important factors on information technology in development free zone markets: An AHP implementation. Management Science Letters, 2(4), 1359-1364.

Babbie, E. (2010). Unobtrusive research. In The practice of social research (pp. 331-361): Wadsworth, Cengage Learning, Belmont, CA.

Babbie, E. R. (1973). Survey research methods: Wadsworth.

Babbie, E. R. (1989). The practice of social research: Wadsworth Publishing Company.

Baindur, D., & Viegas, J. M. (2011). An agent based model concept for assessing modal share in inter-regional freight transport markets. Journal of Transport Geography, 19(6), 1093-1105. doi:https://doi.org/10.1016/j.jtrangeo.2011.05.006

Balcik, B., Beamon, B., & Smilowitz, K. (2008). Last Mile Distribution in Humanitarian Relief. Journal of Intelligent Transportation Systems, 12(2), 51-63. doi:10.1080/15472450802023329

201

Ballantyne, E. E. F., Lindholm, M., & Whiteing, A. (2013). A comparative study of urban freight transport planning: addressing stakeholder needs. Journal of Transport Geography, 32, 93-101. doi:https://doi.org/10.1016/j.jtrangeo.2013.08.013

Banaitiene, N., Banaitis, A., Kaklauskas, A., & Zavadskas, E. K. (2008). Evaluating the life cycle of a building: A multivariant and multiple criteria approach. Omega, 36(3), 429- 441.

Banister, D. (2008). The sustainable mobility paradigm. Transport policy, 15(2), 73-80.

Banker, S. (2016). The Ubernization of last mile freight https://www.forbes.com/sites/stevebanker/2016/04/08/the-uberization-of-last-mile- freight/#d9e72ac3ad49

Banville, C., Landry, M., Martel, J. M., & Boulaire, C. (1998). A stakeholder approach to MCDA. Systems Research and Behavioral Science, 15(1), 15-32. doi:10.1002/(SICI)1099-1743(199801/02)15:1<15::AID-SRES179>3.0.CO;2-B

Baptista, S. R., (2014) Design and use of composite indices in assessments of climate change vulnerability and resilience, USAID. http://www.ciesin.org/documents/Design_Use_of_Composite_Indices.pdf

Barbara, F., Michiya, M., & Jose, M. (Eds.). (2011). Managing Global Supply Chain Relationships: Operations, Strategies and Practices. Hershey, PA, USA: IGI Global.

Barceló, J., Grzybowska, H., & Pardo, S. (2005). Combining vehicle routing models and microscopic traffic simulation to model and evaluating city logistics applications. Paper presented at the proceedings of the 16th mini-EURO conference and 10th meeting of EWGT, Italy.

Baruch, Y. (1999). Response rate in academic studies-A comparative analysis. Human relations, 52(4), 421-438.

Beatley, T. (1995). Planning and sustainability: The elements of a new (improved?) paradigm. CPL bibliography, 9(4), 383-395.

Behrends, S. (2011). Urban freight transport sustainability - the interaction of urban freight and intermodal transport. In.

Bektas, T., Crainic, T. G. and Woensel, T. V. (2016). From Managing Urban Freight to Smart City Logistics Networks, in Pardalos, PM, & Gakis, K (eds) 2016, Network Design and Optimization for Smart Cities, World Scientific Publishing Co Pte Ltd, Singapore.

Bender, A., Din, A., Hoesli, M., & Brocher, S. (2000). Environmental preferences of homeowners: further evidence using the AHP method. Journal of Property Investment & Finance, 18(4), 445-455.

202

Berry, B. J. L., & Garrison, W. L. (1958). A Note on Central Place Theory and the Range of a Good. Economic Geography, 34(4), 304-311. doi:10.2307/142348.

BESTUFS II. (2007). Deliverable D2.3, Part II Best Practice Update 2007 / II, Updated Handbook from Year 2003 Part II Intelligent Transport Systems (ITS). Bhagwat, P., Bhattacharya, P., Krishna, A. & Tripathi, S. K. (1996) Enhancing throughput over wireless LANs using channel state dependent packet scheduling. Proceedings of IEEE INFOCOM - Conference on Computer Communications. 10.1109/INFCOM.1996.493057

Bjorklund, M. & Johansson, H. (2018). Urban consolidation centre – a literature review, categorisation, and a future research agenda, International Journal of Physical Distribution & Logistics Management, 48, 8, 745-764.

Black, J., Hashimzade, N., & Myles, G. (2017). Transaction Cost Economics. In.

Black, K. (2009). Business statistics: Contemporary decision making: John Wiley & Sons.

Blackstone, J. H. (2001). Theory of Constraints-A Status Report.

Blair, N., Hine, J., & Bukhari, S. M. A. (2013). Analysing the impact of network change on transport disadvantage: a GIS-based case study of Belfast. Journal of Transport Geography, 31, 192-200.

Bliemer, M. C. J., Mulley, C., and Moutou, C. J. (2016) Handbook on Transport and Urban Planning in the Developed World. Cheltenham, United Kingdom: Edward Elgar.

Blind, K., 2004. The Economics of Standards: Theory, Evidence, Policy. Edward Elgar: Cheltenham.

Boerkamps, J. K. H., van Binbergen, A. & Bovy, P. H. L., (2000). Modeling Behavioral Aspects of Urban Freight Movement in Supply Chains, Transportation Research Record. Journal of the Transportation Research Board 1725(1725):17-25.

Bogdan, R. C. and Biklen, S. K. (2007). Qualitative Research for Education: An Introduction to Theory and Methods. 5th Edition, Allyn & Bacon, Boston.

Booysen, F. (2002). An overview and evaluation of composite indices of development. Social Indicators Research, 59(2), 115-151. Retrieved from http://dx.doi.org/10.1023/A:1016275505152

Borjeson, L., Hojer, M., Dreborg, K.-H., Ekvall, T., & Finnveden, G. (2006). Scenario types and techniques: towards a user’s guide. Futures, 38(7), 723–739.

Bowen, J. T., (2012). A spatial analysis of FedEx and UPS: hubs, spokes, and network structure. Journal of Transport Geography. 24, 419–431.

Bowersox, D. J., Closs, D. J., & Stank, T. P. (2000). Ten mega-trends that will revolutionize supply chain logistics. Journal of business logistics, 21(2), 1.

203

Bowyer, D. P., & Ogden, K. W. (1988). Impacts of urban road investment on road freight transport: Gateway Bridge case study. Paper presented at the Australian Road Research Board (ARRB) Conference, 14th, 1988, Canberra.

Boyer, K. K., Prud'homme, A. M., & Chung, W. (2009). The last mile challenge: evaluating the effects of customer density and delivery window patterns. Journal of business logistics, 30(1), 185-201.

Boyonas, M. I., Olavarria, L., and Saenz, M. J. (2016) Scenario Planning: Preparing the Future of Global Supply Chains Metro Pedal Power, http://metropedalpower.com.

Bozzo, R., Conca, A., & Marangon, F. (2014). Decision support system for city logistics: literature review, and guidelines for an ex-ante model. Transportation Research Procedia, 3, 518-527.

Bradfield, R., Wright, G., Burt, G., Cairns, G., & Van Der Heijden, K. (2005). The origins and evolution of scenario techniques in long range business planning. Futures, 37(8), 795-812. doi:https://doi.org/10.1016/j.futures.2005.01.003

Brans, J.-P., Vincke, P., & Mareschal, B. (1986). How to select and how to rank projects: The PROMETHEE method. European Journal of Operational Research, 24(2), 228-238.

Breheny, M. J. (1992). Sustainable development and urban form (Vol. 2): Pion London.

Breheny, M., Gordon, I., & Archer, S. (1998). Building densities and sustainable cities. Project Outline, 5.

Browne, M., Allen, J., Steele, S., Cherrett, T., & McLeod, F. (2010). Analysing the results of UK urban freight studies. Procedia-Social and Behavioral Sciences, 2(3), 5956-5966.

Browne, M., Piotrowska, M., Woodburn, A., & Allen, J. (2007). Literature review WM9: Part I-Urban freight transport. Green Logistics Project. University of Westminster, London.

Browne, M., Sweet, M., Woodburn, A., & Allen, J. (2005). Urban freight consolidation centres final report. Transport Studies Group, University of Westminster, 10.

Bryman, A., & Bell, E. (2015). Business research methods: Oxford University Press, USA.

Bukh, P. N. D. (1993). Integral Logistic Structures, Developing Customer-Oriented Goods Flow (Book Review). In (Vol. 23, pp. 133-134).

Burt, G. (2006). Pre-determined elements in the business environment: Reflecting on the legacy of Pierre Wack. Futures, 38(7), 830-840.

Burton, E. (2000). The compact city: Just or just compact? A preliminary analysis. Urban studies, 37(11), 1969-2006. doi:10.1080/00420980050162184

204

Bureau of Infrastructure, Transport and Regional Economics, (2013a), Australian Infrastructure Statistics Yearbook 2013, Statistical Report, BITRE, Canberra. Burell, G. & Morgan, G. (2005). Sociological Paradigms and Organisational Analysis Elements of the Sociology of Corporate Life. Ashgate. Burton, E., Jenks, M., & Williams, K. (2003). The compact city: a sustainable urban form? : Routledge.

Byrne, B. M. (2001). Structural equation modeling: Perspectives on the present and the future. International Journal of Testing, 1(3-4), 327-334.

Cardenas, I., Borbon-Galvez, Y., Verlinden, T., Van de Voorde, E., Vanelslander, T., & Dewulf, W. (2017). City logistics, urban goods distribution and last mile delivery and collection. Competition and Regulation in Network Industries, 18(1-2), 22-43. doi:10.1177/1783591717736505

Casey, N., Rao, D., Mantilla, J., Pelosi, S., & Thompson, R. G. (2014). Understanding Last Kilometre Freight Delivery in Melbourne's Central Business District. Procedia-Social and Behavioral Sciences, 125, 326-333.

Chatterjee, P., Athawale, V. M., & Chakraborty, S. (2011). Materials selection using complex proportional assessment and evaluation of mixed data methods. Materials & Design, 32(2), 851-860.

Chauhan, G. M. S., & Anbalagan, C. (2015). Strategy and Trends in B2B Business: Opportunities and Challenges-A Global Prospective. PDF). International Faculty and Student Multi-Research Consortium. Retrieved, 3.

Chen, L., & Miller-Hooks, E. (2012). Resilience: An Indicator of Recovery Capability in Intermodal Freight Transport. Transportation science, 46(1), 109-123. doi:10.1287/trsc.1110.0376

Chermack, T. J., Lynham, S. A., & van der Merwe, L. (2006). Exploring the relationship between scenario planning and perceptions of learning organization characteristics. Futures, 38(7), 767-777.

Cherrett, T., Allen, J., McLeod, F., Maynard, S., Hickford, A., & Browne, M. (2012). Understanding urban freight activity–key issues for freight planning. Journal of Transport Geography, 24, 22-32.

Chhetri, P., Han, J. H., Chandra, S., & Corcoran, J. (2013). Mapping urban residential density patterns: Compact city model in Melbourne, Australia. City, Culture and Society, 4(2), 77-85. doi:https://doi.org/10.1016/j.ccs.2013.03.001

Chopra, S., & Meindl, P. (2007). Supply chain management: strategy, planning, and operation. Upper Saddle River, N.J.: Pearson/Prentice Hall.

205

Chrisman, N. R. (1998). Rethinking levels of measurement for cartography. Cartography and Geographic Information Systems, 25(4), 231-242.

Chrisman, N. R (1999). What does ‘GIS’ Mean? Transactions in GIS. 3(2): 175±186

Christopher, M. (2016). Logistics & supply chain management: Pearson UK.

City of Boston. Downtown Crossing Vehicular Regulations. http://www.cityofboston.gov/bra/pdf/PlanningPublications/DowntownCrossingVehicul ar Regulations.pdf.

Clercq, F., & de Vries, J. (2000). Public Transport and the Compact City. Transportation Research Record: Journal of the Transportation Research Board, 1735, 3-9. doi:10.3141/1735-01

Cloke, P. J., Philo, C., & Sadler, D. (1991). Approaching human geography an introduction to contemporary theoretical debates.

Coase, R. H. (1937). The Nature of the Firm. Economica, 4(16), 386-405. doi:doi:10.1111/j.1468-0335.1937.tb00002.x

Corbetta, P. (2003) Social Research: Theory, Methods and Techniques. Sage.

Comi, A., & Conte, E. (2011). A modelling system for estimating freight quantities attracted by cities. WIT transactions on the built environment, 423-434.

Conway, A. J., Thuillier, O., Dornhelm, E., & Lownes, N. E. (2013). Commercial vehicle- bicycle conflicts: A growing urban challenge. Paper presented at the Transportation Research Board 92nd Annual Meeting.

Crainic, T. G., Errico, F., Rei, W., & Ricciardi, N. (2015). Modeling Demand Uncertainty in Two-Tier City Logistics Tactical Planning (Vol. 50).

Crainic, T. G., Gendreau, M., & Potvin, J.-Y. (2009). Intelligent freight-transportation systems: Assessment and the contribution of operations research. Transportation Research Part C: Emerging Technologies, 17(6), 541-557.

Crainic, T. G., & Montreuil, B. (2016). Physical internet enabled Hyperconnected City logistics. Transportation Research Procedia, 12, 383-398.

Crainic, T. G., Ricciardi, N., & Storchi, G. (2009). Models for evaluating and planning city logistics systems. Transportation science, 43(4), 432-454.

Crawford, J. H. (2000) Carfree Cities. Utrecht: International Books.

Creswell J. W. (2013). Qualitative Inquiry and Research Design: Choosing Among Five Approaches, SAGE Publications

Creswell, J. W., & Poth, C. N. (2017). Qualitative inquiry and research design: Choosing among five approaches. Sage publications.

206

Crotty, M. (1998). The foundations of social research: Meaning and perspective in the research process: Sage.

Cruijssen, F., Cools, M., & Dullaert, W. (2007). Horizontal cooperation in logistics: Opportunities and impediments. Transportation Research Part E: Logistics and Transportation Review, 43(2), 129-142. doi:https://doi.org/10.1016/j.tre.2005.09.007

Cruijssen, F.; Braysy, O.; Dullaert, W. and Salomon, M. (2007) Joint route planning under varying market conditions. Int. J. Phys. Distribution Logistics Management 37, 287– 304.

Dablanc, L. (2007). Goods transport in large European cities: Difficult to organize, difficult to modernize. Transportation Research Part A: Policy and Practice, 41(3), 280-285.

Dablanc, L. (2008). Urban goods movement and air quality policy and regulation issues in European cities. Journal of Environmental Law, 20(2), 245-266.

Danermark, B. (2002). Explaining society: critical realism in the social sciences. London, New York: Routledge.

Danielis, R., Maggi, E., Rotaris, L., & Valeri, E. (2013). Urban Freight Distribution: Urban Supply Chains and Transportation Policies: Emerald Group Publishing Limited.

Dantzig, G. B., & Saaty, T. L. (1973). Compact city: a plan for a liveable urban environment: WH Freeman.

Dasburg, N., & Schoemaker, J. (2008). Quantification of Urban Freight Transport Effects II. BESTUFS Consortium. In.

Davison, G. (2004) Car wars: How the car won our hearts and conquered our cities, Allen & Unwin,

De Almeida Correia, G. H., & van Arem, B. (2016). Solving the User Optimum Privately Owned Automated Vehicles Assignment Problem (UO-POAVAP): A model to explore the impacts of self-driving vehicles on urban mobility. Transportation Research Part B: Methodological, 87, 64-88.

De Smith, M., Goodchild, M. F., and Longley, P. A. (2009). Geospatial Analysis - A Comprehensive Guide to Principles, Techniques and Software Tools, 3rd edition,

Porta, D., & Keating, M. (2008). Approaches and methodologies in the social sciences: A pluralist perspective: Cambridge University Press.

Denzin, N. K., & Lincoln, Y. S. (2005). The Sage handbook of qualitative research: Sage Publications Ltd.

Department of Infrastructure (2002). Melbourne 2030 Planning for Sustainable Growth, Victoria Government. Department of Infrastructure (2011). Our Cities, Our Future. State Government of Victoria.

207

Department of Planning and Community Development (2008a). Victoria in Future 2008: Information Booklet on Melbourne, 2006–2026. Melbourne: State Government of Victoria. Department of Planning and Community Development (2008b). Melbourne 2030: A planning update Melbourne @ 5 million. Available from: . Deshpande, R. (1983). "Paradigms lost": On theory and method in research in marketing. Journal of Marketing, 47(4), 101-110.

Dettmer, H. W. (2007). The Logical Thinking Process: A Systems Approach to Complex Problem Solving: ASQ Quality Press.

Diziain, D., Taniguchi, E., & Dablanc, L. (2014). Urban Logistics by Rail and Waterways in France and Japan. Procedia - Social and Behavioral Sciences, 125, 159-170. doi:https://doi.org/10.1016/j.sbspro.2014.01.1464

Dolan, S (2018). The challenges of last mile logistics & delivery technology solutions, Business Insider, Australia

Dreborg, K. H. (1996) Essence of Backcasting, Futures, 38(9) pp 813 – 828.

Dueker, K. J., & Ton, T. (2000). Geographical information systems for transport. Handbook of Transport Modelling, 1, 253.

Durance, P., & Godet, M. (2010). Scenario building: Uses and abuses. Technological Forecasting and Social Change, 77(9), 1488-1492. doi:https://doi.org/10.1016/j.techfore.2010.06.007

Dyer, J. S., Fishburn, P. C., Steuer, R. E., Wallenius, J., & Zionts, S. (1992). Multiple criteria decision making, multiattribute utility theory: the next ten years. Management science, 38(5), 645-654.

Ehmke, J. (2012). Integration of information and optimization models for routing in city logistics (Vol. 177): Springer Science & Business Media.

Eppell, V., McClurg, B. A., & Bunker, J. M. (2001). A four level road hierarchy for network planning and management.

Eric, M. Koojiman, S. and Duchâteau, H. (2007). Assessing and Mapping Urban Freight Distribution Initiatives. In Land Use and Transport: European Research Towards Integrated Policies, edited by Stephen Marshall and David Banister, 245-274. Amsterdam: Elsevier.

Esper, T. L., Jensen, T. D., Turnipseed, F. L. and Burton, S. (2003). The Last Mile: An Examination of Effects of Online Retail Delivery Strategies on Consumers. Journal of Business Logistics, 24: 177-203. doi:10.1002/j.2158-1592.2003.tb00051.x

208

Ewedairo, K, Chhetri, P. and Dodson, J. (2015). “A GIS methodology for estimating the transport network impedance to last-mile delivery” 7th State of Australian Cities Conference, 9-11 December, Gold Coast, Australia.

Ewedairo, K, Chhetri, P. and Jie, F. (2016). “Estimating last-mile delivery impedance: a city logistics challenge”, 21st International Symposium on Logistics (ISL), 3-6 July, Kaohsiung, Taiwan.

Ewedairo, K., Chhetri, P., & Jie, F. (2018). Estimating transportation network impedance to last-mile delivery. International Journal of Logistics Management, 29(1), 110-130. doi:10.1108/IJLM-10-2016-0247

Ewedairo, K, Chhetri, P. Dodson, J and Rahman S. (2018). “Building Last Mile Delivery Scenarios: A Case Study of Melbourne”, International Conference on Industrial Engineering and Engineering Management (IEEM), 16 - 19 December, Bangkok, Thailand.

Ewing, R. (1997). Is Los Angeles-style sprawl desirable? Journal of the American planning association, 63(1), 107-126.

Fainstein, S. S., & Campbell, S. (2002). Readings in urban theory: Wiley-Blackwell.

Ferdinand, F. N., Chung, K. H., Ko, H.J., Ko, C. S. (2008) A compromised decision making model for implementing a strategic alliance in express courier services. INF. Int. Interdiscipline. J. 2013, 15, 6173–6188.

Fernandez-Barcelo, I. & Campos-Cacheda, J. M.( 2012). Estimate of social and Environmental costs for the Urban Distribution of Goods. Practical Case for the City of Barcelona. Procedia - Social and Behavioural Sciences, 39, 818-830.

Ferreira, L., Marquez, L., & Smith, N. (2001). Assessing Road-Based Transport Impacts of E-Business.

Figliozzi, M. A. (2009). Planning approximations to the average length of vehicle routing problems with time window constraints. Transportation Research Part B: Methodological, 43(4), 438-447.

Filstead, W. J. (1979). Qualitative methods: A needed perspective in evaluation research. Qualitative and quantitative methods in evaluation research, pp.33-48.

Forer, P. (1993). An implementation of a discrete 3-D space time model of urban accessibility. Paper presented at the Proceedings of the Seventeenth New Zealand Geographical Society Conference.

Forkenbrock, D., & Foster, N. (1997). Accident cost saving and highway attributes. Transportation (1986-1998), 24(1), 79.

209

Freeman, R. E., & Reed, D. L. (1983). Stockholders and stakeholders: A new perspective on corporate governance. California management review, 25(3), 88-106.

Freight, C. (2002). Inter-and intra-city freight-distribution networks. Final report Work package, 1.

Freightline 1. (2013). Australian freight transport overview, Bureau of Infrastructure, Transport and Regional Economics.

Ghaffar, A. A., Alberding, K., Nicholsky, V., and Lai, A., (2005). Scenario Thinking Vision 2020, St Andrew University Press, UK.

Gaigné, C., Riou, S., & Thisse, J.-F. (2012). Are compact cities environmentally friendly? Journal of Urban Economics, 72(2-3), 123-136. doi:10.1016/j.jue.2012.04.001

Galster, G., Hanson, R., Ratcliffe, M. R., Wolman, H., Coleman, S., & Freihage, J. (2001). Wrestling sprawl to the ground: defining and measuring an elusive concept. Housing policy debate, 12(4), 681-717.

Gammelgaard, B. (2015). The emergence of city logistics: the case of Copenhagen’s Citylogistik-kbh. International Journal of Physical Distribution & Logistics Management, 45(4), 333-351.

Gerring, J. (2004). What is a case study and what is it good for? American political science review, 98(2), 341-354.

Getis, A., & Getis, J. (1966). Christaller's Central Place Theory. Journal of Geography, 65(5), 220-226. doi:10.1080/00221346608982415

Gevaers, R., Van de Voorde, E., & Vanelslander, T. (2014). Cost modelling and simulation of last-mile characteristics in an innovative B2C supply chain environment with implications on urban areas and cities. Procedia-Social and Behavioral Sciences, 125, 398-411.

Geyer, R., Van Wassenhove, L. N., & Atasu, A. (2007). The Economics of Remanufacturing under Limited Component Durability and Finite Product Life Cycles. Management science, 53(1), 88-100. doi:10.1287/mnsc.1060.0600

Ghoshal, S., & Moran, P. (1996). Bad for Practice: A Critique of the Transaction Cost Theory. Academy of Management Review, 21(1), 13-47. doi:10.5465/amr.1996.9602161563

Giannopoulos, G. A. (2009). Towards a European ITS for freight transport and logistics: results of current EU funded research and prospects for the future. European Transport Research Review, 1(4), 147-161. doi:10.1007/s12544-009-0022-5

Giannopoulos, G. A. (2016). Urban logistics issues in the European experience. RMIT Workshop on: Urban Freight transport and Logistics.

210

Ginevičius, R., & Podvezko, V. (2009). Evaluating the changes in economic and social development of Lithuanian counties by multiple criteria methods. Technological and Economic Development of Economy, 15(3), 418-436.

Giuliano, G. (2012). The Urban Transport System. Urban Transport and the Environment: An International Perspective, edited by Hideo Nakamura, Yoshitsugu Hayashi, and Anthony D. May, 38-98. Amsterdam: Elsevier.

Giuliano, G., Board, N. R. C. T. R., Program, N. C. F. R., Research, U. S. DoT., & Administration, I. T. (2013). Synthesis of Freight Research in Urban Transportation Planning: Transportation Research Board.

Glaser, B. G. a., & Strauss, A. L. (2017). The discovery of grounded theory: strategies for qualitative research (First edition. ed.). London, New York: Routledge.

Godet, M. (1990). Integration of scenarios and strategic management: Using relevant, consistent and likely scenarios. Futures, 22(7), 730-739. doi:https://doi.org/10.1016/0016-3287(90)90029-H

Goldman, T., & Gorham, R. (2006). Sustainable urban transport: Four innovative directions. Technology in society, 28(1), 261-273.

Goldratt, E. M. (1990). What is this thing called the Theory of Constraints? North River Press, Croton-on-Hudson, NY (1990)

Goldratt, E. M. & Cox, J. (1992). The goal: A process of continuous improvement, North River Press

Gonzalez-Feliu, J., Ambrosini, C., Pluvinet, P., Toilier, F., & Routhier, J.-L. (2012). A simulation framework for evaluating the impacts of urban goods transport in terms of road occupancy. Journal of Computational Science, 3(4), 206-215.

Goodchild, M. F. (2005). GIS and modeling overview. GIS, spatial analysis, and modeling. ESRI Press, Redlands, 1-18.

Goodman, R. (2005). Whatever you call it, just don’t think of last-mile logistics, last. Global Logistics & Supply Chain Strategies, 9(12), 46-51.

Gordon, E., & e Silva, A. d. S. (2011). Net locality: Why location matters in a networked world: John Wiley & Sons.

Gordon, P., & Richardson, H. W. (1997). Are compact cities a desirable planning goal? Journal of the American planning association, 63(1), 95-106.

Goyder, J. (1987). The silent minority: Nonrespondents on sample surveys: Westview Press.

Graham, S., & Marvin, S. (2001). Splintering Urbanism: Networked Infrastructures, Technological Mobilities and the Urban Condition: Routledge.

211

Grant, R. M. (2003). Strategic Planning in a Turbulent Environment: Evidence from the Oil Majors. Strategic Management Journal, 24: 491 – 517.

Gray, D. E. (2013). Doing research in the real world: Sage.

Greasley, A., & Assi, A. (2012). Improving "last mile" delivery performance to retailers in hub and spoke distribution systems. Journal of Manufacturing Technology Management, 23(6), 794-805. doi:10.1108/17410381211253344

Green, G. M., & Wheeler, M. (2004). Awareness and Action in Critical Moments. Negotiation Journal, 20(2), 349-364. doi:10.1111/j.1571-9979.2004.00028.x

Greenbaum, T. (2003). Focus group research: why the traditional research methodology works so effectively and why it deserves to be the most respected of all qualitative research tools. Quirk’s Marketing Research Review, 1-5.

Greeuw, S. C., van Asselt, M. B., Grosskurth, J., Storms, C., Rijkens-Klomp, N., Rothman, D. S., . . . Ribeiro, T. (2000). Cloudy crystal balls. European Environment Agency, Copenhagen, Denmark.

Guarte, J. M., & Barrios, E. B. (2006). Estimation under purposive sampling. Communications in Statistics—Simulation and Computation®, 35(2), 277-284.

Guba, E. G. (1990). The paradigm dialog. Paper presented at the Alternative Paradigms Conference, Mar, 1989, Indiana U, School of Education, San Francisco, CA, US.

Guba, E. G., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research. Handbook of qualitative research, 2(163-194), 105.

Guitouni, A., & Martel, J.-M. (1998). Tentative guidelines to help choosing an appropriate MCDA method. European Journal of Operational Research, 109(2), 501-521.

Gülgen, F., & Gökgöz, T. (2011). A block-based selection method for road network generalization. International Journal of Digital Earth, 4(2), 133-153. doi:10.1080/17538947.2010.489972

Haining, R. P. (2003). Spatial data analysis: theory and practice: Cambridge University Press.

Hajkowicz, S., & Higgins, A. (2008). A comparison of multiple criteria analysis techniques for water resource management. European Journal of Operational Research, 184(1), 255-265.

Han, L., Chin, S.-M., Franzese, O., & Hwang, H. (2005). Estimating the impact of pickup- and delivery-related illegal parking activities on traffic. Transportation Research Record: Journal of the Transportation Research Board (1906), 49-55.

Handfield, R. B. (2011). Sourcing and supply chain management. Andover: South-Western Cengage Learning.

212

Handy, S. L., & Niemeier, D. A. (1997). Measuring accessibility: an exploration of issues and alternatives. Environment and planning A, 29(7), 1175-1194.

Harrington, D. (2015). From First Mile to Last Mile Global Industrial & Logistics Trends, Colliers International.

Harrington, T. S. Srai, J. S., Kumar, M. & Wohlrab, J. (2016) Identifying design criteria for urban system ‘last-mile’ solutions – a multi-stakeholder perspective, Production Planning & Control, 27:6, 456-476, DOI: 10.1080/09537287.2016.1147099

Harris, C D. and Ullman, E. L. (1945). "The Nature of Cities". The Annals of the American Academy of Political and Social Science. 242: 7–17.

He, N and Chao S (2014) Urban Road Traffic Impedance Function——Dalian City Case Study, Journal of Highway and Transportation Research and Development, Vol. 8 Issue (3): 90−95

Healy, M., & Perry, C. (2000). Comprehensive criteria to judge validity and reliability of qualitative research within the realism paradigm. Qualitative Market Research: An International Journal, 3(3), 118-126. doi:doi:10.1108/13522750010333861

Heide, J. B., & Stump, R. L. (1995). Performance implications of buyer-supplier relationships in industrial markets: a transaction cost explanation. Journal of Business Research, 32(1), 57-66.

Hesse, M (2002) Location matters, Access 21, 2 – 26.

Hesse, M., & Rodrigue, J.-P. (2004). The transport geography of logistics and freight distribution. Journal of Transport Geography, 12(3), 171-184.

Hicks, S. (1977) “Urban freight” In: Hensher, D. A. (Ed.) Urban transport economics. Cambridge, UK: Cambridge University Press

Hirsch, R. M., Slack, J. R. and Smith, R.A., 1982. Techniques of trend analysis for monthly water quality data, Water Resources, 18, 107–121

Hoekstra, S., Romme, J., & Argelo, S. M. (1992). Integral logistic structures: developing customer-oriented goods flow. London; New York: McGraw-Hill.

Hoffmann, T., & Prause, G. (2018). On the Regulatory Framework for Last-Mile Delivery Robots. Machines, 6(3), 33.

Hollins, B. (1999). ‘‘Brainstorming products for the long-term future’’, Creativity and Innovation Management, Vol. 8 No. 4, pp. 286-94.

Hubbard, P. (2002). Taking world cities literally: Marketing the city in a global space of flows AU - Doel, Marcus. City, 6(3), 351-368. doi:10.1080/1360481022000037779

213

Huisman, O., & Forer, P. (1998). Computational agents and urban life spaces: a preliminary realisation of the time-geography of student lifestyles. Paper presented at the Proceedings of the 3rd International Conference on GeoComputation.

Hwang, C.-L., & Masud, A. S. M. (2012). Multiple objective decision making—methods and applications: a state-of-the-art survey (Vol. 164): Springer Science & Business Media.

Infrastructure Australia. (2011). National Land Freight Strategy Discussion Paper, Australian Government. Jaller, M., Holguin-Veras, J., & Hodge, S. (2013). Parking in the City: Challenges for Freight Traffic. Transportation research record (2379), 46-56. doi:10.3141/2379-06

Jenks, M., Kozak, D., & Takkanon, P. (2008). World cities and urban form: fragmented, polycentric, sustainable? Abingdon, Oxon; New York: Routledge.

Jiang, B., & Claramunt, C. (2004). A Structural Approach to the Model Generalization of an Urban Street Network*. GeoInformatica, 8(2), 157-171. doi:10.1023/B:GEIN.0000017746.44824.70

Jiang, B., & Harrie, L. (2004). Selection of Streets from a Network Using Self‐Organizing Maps. Transactions in GIS, 8(3), 335-350. doi:10.1111/j.1467-9671.2004.00186.x

Jiang, L., & Mahmassani, H. (2014). City Logistics: Freight Distribution Management with Time-Dependent Travel Times and Disruptive Events. Transportation Research Record: Journal of the Transportation Research Board (2410), 85-95.

Johnson, M. P. (2005). Spatial decision support for assisted housing mobility counselling. Decision Support Systems, 41(1), 296-312.

Joshi, S., & Joshi, R. (2016). Designing and implementing global supply chain management. Hershey, PA, USA: Business Science Reference, an imprint of IGI Global.

Jopson, A (2007).A Land-Use – Transport Vision. Land Use and Transport: European Research Towards Integrated Policies, edited by Stephen Marshall and David Banister, 133-149. Amsterdam: Elsevier.

Jungermann, H., & Thüring, M. (1987). The use of mental models for generating scenarios.

Jurewicz, C., & Excel, R. (2016). Application of a crash-predictive risk assessment model to prioritise road safety investment in Australia. Transportation Research Procedia, 14, 2101-2110.

Juttner, U., Christopher, M., & Baker, S. (2007). Demand chain management-integrating marketing and supply chain management. Industrial Marketing Management, 36(3), 377. doi:10.1016/j.indmarman.2005.10.003

214

Kaklauskas, A., Zavadskas, E. K., Banaitis, A., & Šatkauskas, G. (2007). Defining the utility and market value of a real estate: a multiple criteria approach. International Journal of Strategic Property Management, 11(2), 107-120.

Kaklauskas, A., Zavadskas, E. K., & Raslanas, S. (2005). Multivariant design and multiple criteria analysis of building refurbishments. Energy and Buildings, 37(4), 361-372.

Kaplan, A. (1964). The conduct of inquiry: methodology for behavioral science. Scranton, Pa.: Chandler.

Kawamura, K., Sriraj, P., Surat, H. R., & Menninger, M. (2014). Analysis of Built Environment Features and their Effects on Freight Activities. Procedia-Social and Behavioral Sciences, 125, 28-35.

Keeney, R. L., & Keeney, R. L. (2009). Value-focused thinking: A path to creative decision making: Harvard University Press.

Kenney, M. A., Chen, R. S., Maldonado, J., and Quattrochi, D. (2012). Climate Change Impacts and Responses, NCA Report Series, Volume 5c, Societal Indicators for the National Climate Assessment, April 28-29, 2011. Washington, D.C.: U.S. National Climate Assessment, U.S. Global Change Research Program. Retrieved from http://downloads.usgcrp.gov/NCA/Activities/Societal_Indicators_FINAL.pdf

Kitchin, R., & Tate, N. (2013). Conducting research in human geography: theory, methodology and practice: Routledge.

Knott, R. (1987). The logistics of bulk relief supplies. Disasters, 11(2), 113-115. doi:10.1111/j.1467-7717.1987.tb00624.x

Ko, C. S., Min, H., Ko, H. J. (2007). Determination of cut-off time for express courier services: A Genetic algorithm approach. Int. Trans. Oper. Res. 14, 159–177.

Koshi, M., Yamada, H. and Taniguchi E., 1992, New urban freight transport systems. In: Selected Proceedings of 6th World Conference on Transport Research, Lyon, France, pp. 2117 ± 2128.

Kosow, H., & Gabner, R. (2008). Methods of future and scenario analysis: overview, assessment, and selection criteria.

Krueger, R. A., & Casey, M. A. (2014). Focus groups: A practical guide for applied research: Sage publications.

Lagorio, A., Pinto, R., & Golini, R. (2016). Research in urban logistics: a systematic literature review. International Journal of Physical Distribution & Logistics Management, 46(10), 908-931. doi:doi:10.1108/IJPDLM-01-2016-0008

215

Lai, Kee-hung & Cheng, T. C. E. & C. L. Yeung, Andy. (2005). Relationship stability and supplier commitment to quality. International Journal of Production Economics. 96. 10.1016/j.ijpe.2004.07.005.

Lawrence, P., Meigh, J., and Sullivan, C. (2002). The Water Poverty Index: An international comparison. Keele Economic Research Papers 2002/19. Keele, Staffordshire, UK: Keele University. Retrieved from http://www-docs.tu- cottbus.de/hydrologie/public/scripte/lawrence_etal2002.pdf

Le Clercq, F., & de Vries, J. S. (2000). Public transport and the compact city. Transportation research record, 1835, 3-10.

Lee, S. and Jeong, M. (2012), “Effects of e-servicescape on consumers’ flow experiences”, Journal of Hospitality and Tourism Technology, Vol. 3 No. 1, pp. 47-59.

Levine, D. M., & Stephan, D. F. (2009). Even you can learn statistics: A guide for everyone who has ever been afraid of statistics: FT Press.

Li, C. (2015) Traffic Flow Assignment by Using Road Impedance Function Based on Subjective Preference, International Conference on Chemical, Material and Food Engineering (CMFE-2015)

Li, R., Rose, G and Sarvi M. (2006) Evaluation of Speed-Based Travel Time Estimation Models, Journal of Transportation Engineering, 132(7): 540 – 547.

Li, Z., & Choi, Y. H. (2002). Topographic Map Generalization: Association of Road Elimination with Thematic Attributes. The Cartographic Journal, 39(2), 153-166. doi:10.1179/caj.2002.39.2.153

Lim, S. F. W. T., Lawson, B., & Barratt, M. (2017). Revisiting Theories of Governance for Last-Mile Supply Networks in Omnichannel Retailing. Academy of Management Annual Meeting Proceedings, 2017(1), 1-1. doi:10.5465/AMBPP.2017.12330abstract

Lindawati, Van Schagen, J., Goh, M., & de Souza, R. (2014). Collaboration in urban logistics: motivations and barriers.

Lindgren, M. & Bandhold, H. (2003). Scenario Planning: The link between the future and Strategy. Palgrave Machmilan.

Lindner, J. (2011). Last Mile Logistics Capability: a Multidimensional System Requirements Analysis for a General Modeling and Evaluation Approach. Dipl Technical university of Munich.

Lindholm, M., (2013). Urban freight transport from a local authority perspective – a literature review, European Transport, Issue 54, Paper No. 3, ISSN 1825-3997

Loeppert, P. V., & Boor, S. E. (1995). Impedance circuit for a miniature hearing aid. In: Google Patents.

216

London, T. F. (2012). Delivering a road freight legacy. In: TEAM, F. A. F. (ed.).

López, M. A., Martín, S., Aguado, J. A., & de La Torre, S. (2013). V2G strategies for congestion management in micro grids with high penetration of electric vehicles. Electric Power Systems Research, 104, 28-34. doi:10.1016/j.epsr.2013.06.005

Lö sch, A. (1954). The economics of location, New Haven, Yale University Press.

Lotfi, S., & Solaimani, K. (2009). An assessment of urban quality of life by using analytic hierarchy process approach (case study: comparative study of quality of life in the North of Iran). Journal of Social Sciences, 5(2), 123-133.

Lowe, J. C., (1975). The Geography of Movement. Houghton Mifflin, Boston

Lozano, S.; Moreno, P.; Adenso-Diaz, B.; Algaba, E. (2013) Cooperative game theory approach to allocating benefits of horizontal cooperation. Eur. J. Oper. Res. 229, 444– 452.

Lu, M. & Borbon-Galvez (2012) Advanced logistics and supply chain management for intelligent and sustainable transport, 19th World Congress on ITS.

Luo, Y., Wang, N., Xiang, H., & Wang, J. (2015). Road Impedance Model Study under the Control of Intersection Signal. Mathematical Problems in Engineering, 2015. doi:10.1155/2015/468943

Macharis, C., Milan, L., & Verlinde, S. (2014). A stakeholder-based multicriteria evaluation framework for city distribution. Research in Transportation Business & Management, 11, 75-84. doi:10.1016/j.rtbm.2014.06.004

MacKay, B., & McKiernan, P. (2006). Back to the future: history and the diagnosis of environmental context. International Studies of Management & Organization, 36(3), 93-109.

Mainroads Western Australia (2015). Roundabouts and Traffic Signals: Guidelines for the selection of Intersection Control.

Malczewski, J. (1999). GIS and multicriteria decision analysis: John Wiley & Sons.

Malhotra, Y. (2003). Measuring knowledge assets of a nation: knowledge systems for development. Paper presented at the Invited Research Paper Sponsored by the United Nations Department of Economic and Social Affairs. Keynote Presentation at the Ad Hoc Group of Experts Meeting at the United Nations Headquarters, New York City, NY.

Maliene, V., Grigonis, V., Palevičius, V., & Griffiths, S. (2011). Geographic information system: Old principles with new capabilities. Urban Design International, 16(1), 1-6.

Mann, H. J. (1987) Freight Village (presentation, New York Metropolitan Transportation Council). www.nymtc.org/download_file.cfm?filename=FrtVillage.pdf

217

Mares, R. (2010). The Limits of Supply Chain Responsibility: A Critical Analysis of Corporate Responsibility Instruments, Nordic Journal of International Law, 79,193– 244

Marshall, S. (2005). Streets and Patterns. Spon Press, Oxon and New York.

Mason, R. & Lalwani, C. (2006). Transport integration tools for supply chain management. I. J. of Logistics Research and Applications, 9(1), 57-74.

Mark I., B., Luis, O., & Maria Jesus, S. (2016). Scenario Planning: Preparing the Future of Global Supply Chains. In Designing and Implementing Global Supply Chain Management (pp. 51-91). Hershey, PA, USA: IGI Global.

McDermott, D. R. (1980). Urban Goods Movement: State of the Art and Future Possibilities. Transportation Journal, 34-40.

McGrath, J. E. (1982). Dilemmatics, The Study of Research Choices and Dilemmas, in Judgment Calls in Research (McGrath et al, ed) Sage Publications

McKinnon, A. C. (1999). Vehicle Utilisation and Energy Efficiency in the Food Supply Chain: Full Report of the Key Performance Indicator Survey: Heriot-Watt University, School of Management, Logistics Research Centre.

Melia, S., Parkhurst, G., & Barton, H. (2011). The paradox of intensification. Transport policy, 18(1), 46-52. doi:10.1016/j.tranpol.2010.05.007

Melo, M. T., Nickel, S., & Saldanha-Da-Gama, F. (2009). Facility location and supply chain management – A review. European Journal of Operational Research, 196(2), 401-412. doi:10.1016/j.ejor.2008.05.007

Mendelow, A. and Aubrey L. (1981) Environmental Scanning-The Impact of the Stakeholder Concept, ICIS Conference Proceedings

Meng, Y., Malczewski, J., & Boroushaki, S. (2011). A GIS-based multicriteria decision analysis approach for mapping accessibility patterns of housing development sites: a case study in Canmore, Alberta. Journal of Geographic Information System, 3(01), 50.

Merton, R. (1967). On Theoretical Sociology.

Milestad, R., Svenfelt, A., and Dreborg, K. H. (2014) Developing integrated explorative and normative scenarios: The case of future land use in a climate-neutral Sweden, Futures 60, 59–71

Miller, H. (2008). Transport 2.0: Meeting grand transportation challenges with geographic information science. ArcNews, Winter.

Miller, R. (2007). Futures literacy: A hybrid strategic scenario method. Futures, 39(4), 341- 362.

218

Mills, A. J., Durepos, G., and Wiebe, E. (2010). Encyclopaedia of case study research: L-z; index (Vol. 1): Sage.

Montoya-Torres, J. R., Muñoz-Villamizar, A. and Vega-Mejía, C.A. (2016) On the impact of collaborative strategies for goods delivery in city logistics, Production Planning & Control, 27:6, 443-455, DOI: 10.1080/09537287.2016.1147092

Morgan, D. L., & Krueger, R. A. (1998). Analyzing and reporting focus group results, Vol. 6: Sage.

Morganti, E. (2011). Urban food planning and transport sustainability: A case study in Parma, Italy. Paper presented at the Proceedings of European Association of Agricultural Economists EAAE 4th PhD Workshop.

Morganti E. & Dablanc L. (2014). Recent Innovation in Last Mile Deliveries. In: Non- technological Innovations for Sustainable Transport. Springer Briefs in Applied Sciences and Technology. Springer, Cham

Morganti, E., & Gonzalez-Feliu, J. (2015). The last food mile concept as a city logistics solution for perishable products. Enterprise Interoperability: Interoperability for Agility, Resilience and Plasticity of Collaborations (I-ESA 14 Proceedings), 202.

Morganti, E., Seidel, S., Blanquart, C., Dablanc, L & Lenz, B. (2014). The Impact of E- commerce on Final Deliveries: Alternative Parcel Delivery Services in France and Germany, Transportation Research Procedia, Volume 4, pp. 178-190.

Morris, A. (2009). The Last Mile: Developing Efficient Freight Operations for Manhattan’s Buildings. Stephen L. Newman Real Estate Institute, Baruch College, New York.

Morris, A., Kornhauser, A., and Kay, M. (1999). Getting the goods delivered in dense urban areas: a snapshot of the last link of the supply chain. Transportation Research Record: Journal of the Transportation Research Board (1653), 34-41.

Morris, A. G., Kornhauser, A. L. & Winship, C. Y. (2003). Solving Street Congestion in the “First and Last Mile”: Decreasing Turnaround Time in Deliveries to Commercial Buildings. 82nd Annual Meeting. Transportation Research Board.

Morrison, N. (1998). The compact city: Theory versus practice—The case of Cambridge. Netherlands journal of housing and the built environment, 13(2), 157-179. doi:10.1007/BF02497227

Muijs, D. (2010). Doing quantitative research in education with SPSS: Sage.

Mulcaster, W. (2008). Three Strategic Frameworks, Business Strategy Series, 10. 1, 68-75.

Mulliner, E., Smallbone, K., & Maliene, V. (2013). An assessment of sustainable housing affordability using a multiple criteria decision making method. Omega, 41(2), 270-279.

219

Muñoz-Villamizar, A., Montoya-Torres, J. R., & Vega-Mejía, C. A. (2015). Non- Collaborative versus Collaborative Last-Mile Delivery in Urban Systems with Stochastic Demands. Procedia CIRP, 30, 263-268. doi:10.1016/j.procir.2015.02.147

Muñuzuri, J., Larrañeta, J., Onieva, L., & Cortés, P. (2005). Solutions applicable by local administrations for urban logistics improvement. Cities, 22(1), 15-28.

Murray, S., Bertram, N. Khor, L., Rowe, D., Meyer, B., Murphy, C., Newton, P., Glackin, S., Alves, T. and McGauran, R (2015). Processes for developing affordable and sustainable medium-density housing models for Greyfield Precincts, AHURI.

Naesens, K., Gelders, L., & Pintelon, L. (2009). A swift response framework for measuring the strategic fit for a horizontal collaborative initiative. International Journal of Production Economics, 121(2), 550-561. doi:10.1016/j.ijpe.2007.04.004

Nagy, G., & Salhi, S. Location-routing: Issues, models and methods. European Journal of Operational Research, 177(2), 649.

National Transport Commission (Road Transport Legislation - Australian Road Rules) Regulations (2006) Australian Road Rule. https://www.legislation.gov.au/Details/F2008C00408/Controls/

Natividade-Jesus, E., Coutinho-Rodrigues, J., & Antunes, C. H. (2007). A multicriteria decision support system for housing evaluation. Decision Support Systems, 43(3), 779- 790.

Neuman, M. (2005). The compact city fallacy. Journal of planning education and research, 25(1), 11-26.

Neyman, J. (1937). Outline of a theory of statistical estimation based on the classical theory of probability. Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, 236(767), 333-380. van Notten, Ph. W. F. (2000), “Create the Future: 21-22 June Workshop Report”, ICIS, Maastricht. van Notten, Ph. W. F. (2005), “Writing on the Wall: Scenario Development in Times of Discontinuity”, Dissertation, www.dissertation.com van Notten Ph. W. F. (2006) Scenario Development: A Typology of Approaches. Thinky Scenarios, Rethink Education OECD.

Nulty, D. D. (2008). The adequacy of response rates to online and paper surveys: what can be done? Assessment & Evaluation in Higher Education, 33(3), 301-314.

OECD (2002). Transport Logistics: Shared solutions for common challenges, Organisation for Economic Cooperation and Development, OECD Publishing, pp. 27-34.

220

Guillemette, Y. & Turner. D. (2018), The Long View: Scenarios for the World Economy to 2060, OECD Economic Policy Papers, No. 22, OECD Publishing, Paris, https://doi.org/10.1787/b4f4e03e-en.

Ogden, K. W. (1992). Urban goods movement: a guide to policy and planning.

Oishi, R., 1996, A study on planning methodology for new urban freight transport systems. PhD dissertation, Kyoto University.

O’Kelly, Morton E., (1998). A geographer’s analysis of hub-and-spoke networks. Journal of Transport Geography 6 (3), 171–186.

Oliva, R., and Watson, N. (2011). Cross-functional alignment in supply chain planning: A case study of sales and operations planning. Journal of Operations Management, 29(5), 434-448. doi:https://doi.org/10.1016/j.jom.2010.11.012

Park, H., Park, D., and Jeong, I. J. (2016). An effects analysis of logistics collaboration in last-mile networks for CEP delivery services. Transport policy, 50, 115-125. doi:10.1016/j.tranpol.2016.05.009

Park, M., & Regan, A. (2004). Issues in Emerging Home Delivery Operations. UC Berkeley: University of California Transportation Center. Retrieved from https://escholarship.org/uc/item/3754142p

Patier, D and Browne, M. (2010) A methodology for the evaluation of urban logistics innovations Procedia Social and Behavioral Sciences 2: 6229–6241

Patton, M. Q. (1987). How to use qualitative methods in evaluation. London: Sage.

Payne, G., & Payne, J. (2004). Key concepts in social research: Sage.

Perera, P., Perera, H., & Wijesinghe, T. (2013). Environmental performance evaluation in supply chain. Vision, 17(1), 53-61.

Pivo, G., Carlson, D., Kitchen, M., & Billen, D. (2002). Learning From Truckers: Truck Drivers' views on the Planning and Design of Urban and Suburban Centres. Journal of Architectural and Planning Research, 12-29.

Postma, T. & Liebl, F., (2005). How to improve scenario analysis as a strategic management tool? Technological Forecasting and Social Change, Volume 72, Issue 2.

Porta, D. D. & Keating, M. (2008). Approaches and Methodologies in the Social Sciences: A Pluralist Perspective, Cambridge.

Prajogo, D. I. (2005). The comparative analysis of TQM practices and quality performance between manufacturing and service firms. International Journal of Service Industry Management, 16(3), 217-228.

221

Pronello, C., Camusso, C. & Rappazzo, V. (2017) Last Mile Freight Distribution and Transport Operators’ Needs: Which Targets and Challenges? Transportation Research Procedia, 25, 0

Quak, H. (2011). Urban freight transport: the challenge of sustainability. City distribution and urban freight transport: Multiple perspectives, 37-55.

Rahman, S. U. (1998). Theory of Constraints: A review of the philosophy and its applications. International Journal of Operations & Production Management, 18(4), 336-355. doi:doi:10.1108/01443579810199720

Rahman, S. U. (2002). The Theory of Constraints’ thinking process approach to developing strategies in supply chains. International Journal of Physical Distribution & Logistics Management, 32(10), 809-828. doi:doi:10.1108/09600030210455429

Ramirez, R., Mukherjee, M., Vezzoli, S., & Kramer, A. M., (2015). Scenarios as a scholarly methodology to produce “interesting research”, Futures, Volume 71, pp. 70-87.

Rand, G. K. (2000). Critical chain: the Theory of Constraints applied to project management. International Journal of Project Management, 18(3), 173-177.

Randhir, T. O., O’Connor, R., Penner, P. R., & Goodwin, D. W. (2001). A watershed-based land prioritization model for water supply protection. Forest Ecology and Management, 143(1), 47-56.

Rao, C., Goh, M., Zhao, Y., & Zheng, J. (2015). Location selection of city logistics centres under sustainability. Transportation Research Part D: Transport and Environment, 36, 29-44.

Reilly, W. J. (1953). The law of retail gravitation, New York, NY: Pilsbury Publishers

Reisman, A. and Chase, M. (2011) Strategies for Reducing the Impacts of Last-Mile Freight in Urban Business Districts, Final Paper, UEP-173: Transportation Planning.

Ringland, G. (2006). Scenario planning managing for the future. John Wiley & Sons.

Ringland, G., & Schwartz, P. P. (1998). Scenario planning: managing for the future: John Wiley & Sons.

Road Management Act (2004). No. 12 of 2004. State of Victoria.

Robinson, J., Burch, S., Talwar, S., O’Hea, M., and Walsh, M. (2011). Envisioning sustainability: Recent progress in the use of participatory backcasting approaches for sustainability research, Technological Forecasting and Social Change. 78, 5. Pp. 756 – 768.

Rodrigue, J-P. and Dablanc L., (2017), City Logistics, In The Geography of Transport Systems, Rodrigue J-P., Fourth edition New York.

222

Rodrigue, J-P., Comtois, C., & Slack, B. (2013). The geography of transport systems: Routledge.

Rodrigues, D. S., Ribeiro, P. J. G., & da Silva Nogueira, I. C. (2015). Safety classification using GIS in decision-making process to define priority road interventions. Journal of Transport Geography, 43, 101-110.

Romeo, D., Rotaris, L. and Marcucci, E. (2010) Urban Freight Policies and Distribution Channels.‖ European Transport, no. 46, 114-146. The Sixth International Conference on City Logistics

Rooijena, T. and Quaka, H. (2010) Local impacts of a new urban consolidation centre – the case of Binnenstadservice.nl, Procedia Social and Behavioral Sciences 2, 5967–5979.

Rosenhead, J., & Mingers, J. (2001). Rational analysis for a problematic world revisited: John Wiley and Sons.

Rotmans, J. & Van Asselt, M. B. A. (2000). Towards an integrated approach for sustainable city planning. Journal of Multi- Criteria Decision Analysis, 9: 110-124.

Roy, B. (1990). The outranking approach and the foundations of ELECTRE methods. In Readings in multiple criteria decision aid (pp. 155-183): Springer.

Rumelt, R. P. (2011). Good Strategy Bad Strategy: The difference and Why it Matters. Crown Business, New York.

Russo, F., & Comi, A. (2011). A model system for the ex-ante assessment of city logistics measures. Research in Transportation Economics, 31(1), 81-87. doi:https://doi.org/10.1016/j.retrec.2010.11.011

Saaty, T. L. (1995). Transport planning with multiple criteria: the analytic hierarchy process applications and progress review. Journal of advanced transportation, 29(1), 81-126.

Salas, M. N. J., & Kaplan, A. (1967). The Conduct of Inquiry. Revista española de la opinión pú blica (10), 379. doi:10.2307/40180865

Sandy Edward Green, J. (2004). A Rhetorical Theory of Diffusion. Academy of Management Review, 29(4), 653-669. doi:10.5465/amr.2004.14497653

Sancha, C., Gimenez, C. & Sierra, V. (2016). Achieving a socially responsible supply chain through assessment and collaboration, Journal of Cleaner Production 112, 3, 1934- 1947.

Sarker, B. R. (2002). A review of: “Supply Chain Management: Strategy, Planning and Operation” Sunil Chopra and Peter Meindl Prentice Hall, Inc., 2001, 457 pages, ISBN: 0-13-026465-2. IIE Transactions, 34(2), 221-222. doi:10.1080/07408170208928863

Sarpong, D. (2011). Towards a methodological approach: theorising scenario thinking as a social practice. Foresight, 13(2), 4-17.

223

Sarpong, D. & Maclean, M. (2011). Scenario thinking: A practice based approach for the identification of opportunities for innovation. Futures 43 (10): 1154 – 1163 .

Sarpong, D., Maclean, M. & Davis, C. (2013). Matter of foresight: How organizing practices enable (or impede) organizational foresightfulness. European Management Journal 31 ( 6 ): 613 – 625 .

Sarraj, R., Ballot, E., Pan, S., Hakimi, D., & Montreuil, B. (2014). Interconnected logistic networks and protocols: simulation-based efficiency assessment. International Journal of Production Research, 52(11), 3185-3208.

Schnaars, S. P. (1987). How to develop and use scenarios, Long Range Planning, Volume 20, Issue 1, Pages 105-114

Schmenner, R. W. (2001), Looking Ahead By Looking Back: Swift, Even Flow In The History Of Manufacturing. Production and Operations Management, 10: 87-96. doi:10.1111/j.1937-5956.2001.tb00069.x

Schoemaker, P. (1991). When and How to Use Scenario Planning: A Heuristic Approach with Illustration (Vol. 10).

Schoemaker, J. (2002). Stadsdistributiecentrum leiden. In D. Egger, & M. Ruesch (Eds.), BESTUFS – Best practice handbook 2002. BESTUFS, 1999-TN.10003

Schwartz, Peter, (1991). The Art of the Long View: Planning for the Future in an Uncertain World, New York: Currency Doubleday, p. 7

Schwartz, M. S. & Carroll, A. B. (2003). Corporate Social Responsibility: A Three-Domain Approach, Business Ethics Quarterly 13(4):503-530

Simatupang, T. M., Wright, A. C., & Sridharan, R. (2004). Applying the Theory of Constraints to supply chain collaboration. Supply Chain Management: An International Journal, 9(1), 57-70. doi:doi:10.1108/13598540410517584

Sivakumar, A., & Bhat, C. R. (2007). Comprehensive, Unified Framework for Analyzing Spatial Location Choice. Transportation research record, 2003(1), 103-111. doi:10.3141/2003-13

Smith, M. (1978). Urban goods movement. Paper presented at the Transportation and Traffic Engineering: Proceedings of a Meeting of the Group Held in Conjunction with the Annual Conference of the NZIE, Hamilton.

Smith, N., Ferreira, L., & Marquez, L. (2001). Assessing road-based transport impacts of E- Business. Paper presented at the AustralAsian Transport Research Forum.

Song, L., Cherrett, T., McLeod, F., & Guan, W. (2009). Addressing the last mile problem: transport impacts of collection and delivery points. Transportation Research Record: Journal of the Transportation Research Board (2097), 9-18.

224

Srai, J.S., Harrington, T. S. and Alinaghian, L.S. (2014) An approach to exploring integrationbenefits of continuous flow technologies within Pharmaceutical supply chains. Chimica Oggi - Chemistry Today, 32. pp. 14-19.

Srinivasan, V. C. (1994). Using the analytic hierarchy process in house selection. The Journal of Real Estate Finance and Economics, 9(1), 69-85.

Stake, R. E. (1995). The art of case study research: Sage.

Stank, T. P., & Goldsby, T. J. (2000). A framework for transportation decision making in an integrated supply chain. Supply Chain Management: An International Journal, 5(2), 71-78. doi:10.1108/13598540010319984

Stathopoulos, A., Valeri, E., & Marcucci, E. (2012). Stakeholder reactions to urban freight policy innovation. Journal of Transport Geography, 22, 34-45.

STATISTICS, A. B. O. (2015). Road Freight Movement, Australia. . Retrieved from http://www.abs.gov.au/ausstats/[email protected]/0/7480A96EA1DF4527CA257EEC000EEB5 D?Opendocument:

Stead, D., & Marshall, S. (2001). The relationships between urban form and travel patterns. An international review and evaluation. European Journal of Transport and Infrastructure Research, 1(2), 113-141.

Stewart, T. J. (1992). A critical survey on the status of multiple criteria decision making theory and practice. Omega, 20(5-6), 569-586.

Stock, J. R., & Boyer, S. L. (2009). Developing a consensus definition of supply chain management: a qualitative study. International Journal of Physical Distribution & Logistics Management, 39(8), 690-711. doi:doi:10.1108/09600030910996323

Stokols, D., Novaco, R. W., Stokols, J., & Campbell, J. (1978). Traffic congestion, Type A behavior, and stress. Journal of Applied Psychology, 63(4), 467.

Stratec, S. (2005). City Freight. Inter-and Intra-city Freight Distribution Networks. City Freight Project, European Commission Fifth Framework Programme.

Sudhanshu, J., & Rohit, J. (Eds.). (2016). Designing and Implementing Global Supply Chain Management. Hershey, PA, USA: IGI Global.

Sutton, R. I., & Staw, B. M. (1995). What Theory is Not. Administrative Science Quarterly, 40(3), 371-384. doi:10.2307/2393788

Taniguchi, E. (2001). City logistics. Infrastructure Planning Review, 18, 1-16.

225

Taniguchi, E., Rob E. C. M. and Der Heijden, V. (2000). An evaluation methodology for city logistics, Transport Reviews, 20:1, 65-90, DOI: 10.1080/014416400295347

Taniguchi, E., & Tamagawa, D. (2005). Evaluating city logistics measures considering the behavior of several stakeholders. Journal of the Eastern Asia Society for Transportation Studies, 6, 3062-3076.

Taniguchi, E., & Thompson, R. (2002). Modeling city logistics. Transportation Research Record: Journal of the Transportation Research Board (1790), 45-51.

Taniguchi, E., Thompson, R. G., & Yamada, T. (2014). Recent trends and innovations in modelling city logistics. Procedia-Social and Behavioral Sciences, 125, 4-14.

Temporal, P. (2005). B2B Branding–A Guide to Successful Business-to-Business Brands, International Enterprise Singapore.

Thomas D. J and Griffin, P. M. (1996). Coordinated supply chain management European Journal of Operational Research 94 (1996) 1-15

Thomson, R., & Brooks, R. (2007). Chapter 13 - Generalisation of Geographical Networks A2 - Mackaness, William A. In A. Ruas & L. T. Sarjakoski (Eds.), Generalisation of Geographic Information (pp. 255-267). Amsterdam: Elsevier Science B.V.

Thord, R., Andersson, Å. E., & Thord, R. (1993). The future of transportation and communication: visions and perspectives from Europe, Japan, and the USA: Springer.

Tingle R. (2018) Ocado wants to whisk goods through tunnels with Elon Musk-style electromagnet technology in just three years Daily Mail 28 September 2018.

Tipagornwong, C., & Figliozzi, M. (2014). Analysis of competitiveness of freight tricycle delivery services in urban areas. Transportation Research Record: Journal of the Transportation Research Board (2410), 76-84.

Toth, P., & Vigo, D. (2002). Models, relaxations and exact approaches for the capacitated vehicle routing problem. Discrete Applied Mathematics, 123(1), 487-512. doi:10.1016/S0166-218X(01)00351-1

Transportation Research Board (2000) Highway Capacity Manual, Transportation Research Board, Washington, DC, USA.

Trentini, A., & Malhene, N. (2012). Flow Management of Passengers and Goods Coexisting in the Urban Environment: Conceptual and Operational Points of View. Procedia - Social and Behavioral Sciences, 39, 807-817. doi:https://doi.org/10.1016/j.sbspro.2012.03.149

Triantaphyllou, E. Multi-criteria decision making methods: a comparative study. 2000. Netherlands: Kluwer Academic Publishers.

226

Tschopp, M., Fröhlich, P., & Axhausen, K. W. (2005). Accessibility and spatial development in Switzerland During the last 50 years. In Access to destinations (pp. 361-376): Emerald Group Publishing Limited.

Tsou, C.-M. (2013). On the strategy of supply chain collaboration based on dynamic inventory target level management: A theory of constraint perspective. Applied Mathematical Modelling, 37(7), 5204-5214. doi:https://doi.org/10.1016/j.apm.2012.10.031

Union, E. (2001). Commission notice: guidelines on the applicability of Article 81 of the EC Treaty to horizontal cooperation agreements. EU.

United Nations, (2016). World Population Prospects: The 2016 Revision, Key Findings and Advance Tables. Working Paper No. ESA/P/WP.241. United Nations Department Of Economic and Social Affairs, Population Division (2012) World Urbanisation Prospects: The 2012 Revision.

United Nations Department Of Economic and Social Affairs, Population Division (2014) World Urbanisation Prospects: The 2014 Revision.

Urban Services Design standards for urban infrastructure - 3 road design https://www.tccs.act.gov.au/__data/assets/pdf_file/0011/396857/ds03_roaddesign.pdf

Užšilaityte, L., & Martinaitis, V. (2010). Search for optimal solution of public building renovation in terms of life cycle. Journal of environmental engineering and landscape management, 18(2), 102-110.

Vaaland, T. I & Owusu, R. A. (2012). What is a Responsible Supply Chain? International Journal of Business and Management, 7, 4; F

Vakulenko, Y., Hellström, D., & Hjort, K. (2018). What's in the parcel locker? Exploring customer value in e-commerce last mile delivery. Journal of Business Research, 88, 421-427. doi:https://doi.org/10.1016/j.jbusres.2017.11.033

Van Rooijen, T., & Quak, H. (2010). Local impacts of a new urban consolidation centre–the case of Binnenstadservice. nl. Procedia-Social and Behavioral Sciences, 2(3), 5967- 5979.

Verlinde, S., Macharis, C., & Witlox, F. (2012). How to consolidate urban flows of goods without setting up an urban consolidation centre? Procedia-Social and Behavioral Sciences, 39, 687-701.

VicRoads (2000) Guidelines for providing loading zones and other short term drop off/pick up area. Vicroads publication Number 00417.

Victoria Government(1987) Planning and Environment Act http://www.austlii.edu.au/cgi- bin/download.cgi/cgi- bin/download.cgi/download/au/legis/vic/consol_act/paea1987254.pdf

227

Victoria Planning Provision (2019). Planning Scheme Online, Victorian Government.

Vissper, J. G. S. N., 1997, Underground networks for freight transport: a dedicated infrastructure for intermodal short distance freight transport. In: Proceedings of 3rd TRAIL Year Congress on Transport, Infrastructure and Logistics, Delft, The Netherlands.

Visser, J., Van Binsbergen, A. & Nemoto, T. (1999). Urban freight transport policy and planning. City logistics I, 39-70.

Visser, J., Nemoto, T. & Browne, M. (2014). Home Delivery and the Impacts on Urban Freight Transport: A Review, Procedia - Social and Behavioral Sciences, Volume 125, pp 15-27,

Viteikiene, M., & Zavadskas, E. K. (2007). Evaluating the sustainability of Vilnius city residential areas. Journal of civil engineering and management, 13(2), 149-155.

Wack, P. (1995). Scenarios: uncharted waters ahead. Harvard Business Review September– October.

Wagner, W. B. (1974). An Empirical Test of Reilly's Law of Retail Gravitation. Growth and Change, 5: 30-35. doi:10.1111/j.1468-2257.1974.tb01019.x

Wegener, M., & Fürst, F. (1999). Land-use transport interaction: State of the art. Berichte aus dem Institut für Raumplanung, 46.

Wakefield, S. (2018). Infrastructures of liberal life: From modernity and progress to resilience and ruins. Geography Compass, 12(7), e12377. doi:doi:10.1111/gec3.12377

Walker, G. and Manson, A. (2014). Telematics, urban freight logistics and low carbon road networks, Journal of Transport Geography, 37 74–81.

Walker, G., & Weber, D. (1984). A Transaction Cost Approach to Make-or-Buy Decisions, Administrative Science Quarterly, 29, 3. 373-391.

Wang, Y., Zhang, J., Assogba, K., Liu, Y., Xu, M., & Wang, Y. (2018). Collaboration and transportation resource sharing in multiple centers vehicle routing optimization with delivery and pickup. Knowledge-Based Systems, 160, 296-310. doi:https://doi.org/10.1016/j.knosys.2018.07.024

Watt, S., Simpson, C., McKillop, C., & Nunn, V. (2002). Electronic course surveys: does automating feedback and reporting give better results? Assessment & Evaluation in Higher Education, 27(4), 325-337.

Weber, A. (1909). Theory of the location of industries. Chicago: University of Chicago Press.

Wegener, M., & Fürst, F. (2004). Land-use transport interaction: state of the art.

228

Whiteing, A., Browne, M. and McKinnon, A., (2010) Green Logistics: Improving the Environmental Sustainability of Logistics. London, England: Kogan Page Limited.

Williamson, O. E. (1977). Markets and Hierarchies. Challenge, 20(1), 70-72.

World Bank (2015) World development report 2015: Mind, Society, and Behavior, Wold Bank Publications

Wolpert, S., & Reuter, C. (2012). Status Quo of City Logistics in Scientific Literature: Systematic Review. Transportation Research Record: Journal of the Transportation Research Board (2269), 110-116.

Wongthavarawat, K. Ganz, A. (2003) Packet scheduling for QoS support in IEEE 802.16 broadband wireless access systems, Int. J. Communication System, 16, 81-96.

Worasinchai, L., & Daneshgar, F. (2012). A qualitative analysis of knowledge transfer in global supply chains: Case of Thai distributer of imported products. Electronic Journal of Knowledge Management, 10(2), 195-204.

Wright, G., & Cairns, G. (2011). Scenario thinking: Practical approaches to the future: Springer.

Wu, H. J., & Dunn, S. C. (1995). Environmentally responsible logistics systems. International Journal of Physical Distribution & Logistics Management, 25(2), 20-38. doi:doi:10.1108/09600039510083925

Xu, L., & Yang, J.-B. (2001). Introduction to multi-criteria decision making and the evidential reasoning approach: Manchester School of Management.

Yager, R. R. (1988). On ordered weighted averaging aggregation operators in multicriteria decision making. IEEE Transactions on systems, Man, and Cybernetics, 18(1), 183- 190.

Yatsalo, B. I., Kiker, G. A., Kim, J., Bridges, T. S., Seager, T. P., Gardner, K., … Linkov, I. (2007). Application of multicriteria decision analysis tools to two contaminated sediment case studies. Integrated Environmental Assessment and Management, 3(2), 223-233.

Yin, R. (2009). Case study: Design and methods 4th ed. Paper presented at the United States: Library Congress Data.

Yuan Wanga, (2016). Towards enhancing the last-mile delivery: An effective crowd-tasking model with scalable solutions. Transportation Research Part E, 3, 279-293.

Yulia Vakulenko, D. H., Klas Hjort. (2018). What's in the parcel locker? Exploring customer value in e-commerce last mile delivery Journal of Business Research, 88, 421-427.

229

Zakuan, N., Yusof, S. M., Laosirihongthong, T., & Shaharoun, A. M. (2010). Proposed relationship of TQM and organisational performance using structured equation modelling. Total Quality Management, 21(2), 185-203.

Zanakis, S. H., Solomon, A., Wishart, N., & Dublish, S. (1998). Multi-attribute decision making: a simulation comparison of select methods. European Journal of Operational Research, 107(3), 507-529.

Zopounidis, C. (1999). Multicriteria decision aid in financial management. European Journal of Operational Research, 119(2), 404-415.

230

Appendix A: Ethics Approval

231

INVITATION TO PARTICIPATE IN A RESEARCH PROJECT

PARTICIPANT INFORMATION

Project Title: “Estimating Last Mile Delivery Impedance: a City Logistics Challenge”

Investigators:

(1) Prof. Prem Chhetri, PhD,

(2) Prof. Jago Dodson,, PhD,

(3) Kolawole Ewedairo,

Dear Participant,

You are invited to participate in a research project being conducted by RMIT University. Please read this sheet carefully and be confident that you understand its contents before deciding whether to participate. If you have any questions about the project, please ask one of the investigators.

Who is involved in this research project? Why is it being conducted?

My name is Kolawole Ewedairo and I am doing a PhD research in the School of Business IT & Logistics, RMIT University, Melbourne. My supervisors are Prof. Chhetri and Prof. Dodson. This project has been approved by the RMIT Human Research Ethics Committee. The primary goal of this research is to estimate and map last mile delivery impedance using planning and transport controls to mitigate against delay and increase last mile delivery efficiency.

Why have you been approached?

This survey is to be completed by the logisticians within business to business last mile delivery organisation. If you are a driver or work within city logistics last mile delivery, you are invited to participate in this PhD research project being conducted through RMIT University.

What is the project about? What are the questions being addressed?

We define last mile as the as the final leg in the supply chain management in cities from the business to business perspective. Impedance is defined as the amount of resistance imposed to traverse through a route in a network from the point pick-up to the point of delivery.

Last mile delivery is the most inefficient and most costly within the supply chain management especially within the urban setting. This is due to its complexity and the impact of the built and regulatory environment on last mile delivery. Thus examining how the attributes of the built and regulatory environment impedes last mile delivery to recommend mitigating strategies for efficient and sustainable last mile delivery is important.

232

The participants will be asked to share their knowledge on how the built environment through planning and transport controls impedes last mile delivery.

If I agree to participate, what will I be required to do?

If you decide to participate, you will be asked to complete a questionnaire which will take approximately 30 minutes to complete. The questionnaire can be assessed on line or can be physically completed.

What are the benefits associated with participation?

The built and regulatory environment contributes to the severity of last mile delivery problem in urban areas. The use of land use controls as instruments to increase city compactness in strategic nodes/hubs is more likely deter the movement of freight. The mapped outputs will help urban planners and logisticians in mitigating the potential delay in last mile delivery through devising localised strategies such as dedicated freight corridors or time-bound deliveries in congested areas to improve retail businesses.

What will happen to the information I provide?

The responses you provide to the survey will be stored on RMIT University server. Once we have completed our data collection and analysis, we will import the data we collect to the RMIT server where it will be stored securely for a period of five (5) years. Data will be reported as an aggregate data. Therefore, individuals will not be identified. Your privacy and confidentiality will be strictly maintained in such a manner that you will not be identified in the thesis report or publication. As participants’ details are not recorded, any information that you provide can be disclosed as aggregate data only if (1) it is to protect you or others from harm, (2) if specifically required or allowed by law, or (3) you provide the researchers with written permission. Data will be only seen by the researcher and supervisors who will also protect you from any risks.

At the conclusion of the project, a summary of the results and associated reports will be made available should you request for it. If you wish to receive the results of this study, then please email your contact details [email protected]. The contact details will be used strictly for dissemination of results and will not be passed to third party and will be purged once the objective is met. The final results will also be reported in a thesis to be submitted for Mr. Kolawole Ewedairo PhD degree, and as appropriate, in papers for presentation at conferences or for publication in scientific journals. Because of the nature of data collection, we are not obtaining written informed consent from you. Instead, we assume that you have given consent by your completion and return of the questionnaire.

What are my rights as a participant?

As a participant, you have the right to withdraw at any time and to have any questions answered at any time. Your participation in this research will help to add weight to the secondary data collected on identified built environments attributes

233

What are the possible risks or disadvantages? Whom should I contact if I have any questions?

There are no anticipated risks associated with participation. However, if you are unduly concerned about your responses to any of the questionnaire items or if you find participation in the project distressing, you should contact the Ethics Officer, Research Integrity, Governance and Systems, RMIT University, as soon as convenient. The Ethics Officer will discuss your concerns with you confidentially and suggest appropriate follow-up, if necessary.

Security of the website

Users should be aware that the World Wide Web is an insecure public network that gives rise to the potential risk that a user’s transactions are being viewed, intercepted or modified by third parties or that data which the user downloads may contain computer viruses or other defects.

Security of the data

This project will use an external site to create, collect and analyse data collected in a survey format. The site we are using is https://www.qualtrics.com. If you agree to participate in this survey, the responses you provide to the survey will be stored on a host server that is used by Qualtrics. No personal information will be collected in the survey so none will be stored as data. Once we have completed our data collection and analysis, we will import the data we collect to the RMIT server where it will be stored securely for five (5) years. The data on the Qualtrics host server will then be deleted and expunged.

Thank you for your assistance and for giving us your time to participate. We value your contribution to this research.

Yours sincerely

Kolawole Ewedairo,

Prof. Prem Chhetri

Prof. Jago Dodson.

If you have any concerns about your participation in this project, which you do not wish to discuss with the researchers, then you can contact the Ethics Officer, Research Integrity, Governance and Systems, RMIT University

234

Participant Information Sheet/Consent Form

Estimating last mile delivery impedance: A case Title of Melbourne.

Chief Investigator/Senior Supervisor Professor Prem Chhetri Associate Investigator/Associate Professor Jago Dodson and Prof Paul Lee Supervisor Principal Investigator]

Principal Research Student Kolawole Ewedairo

What does my participation involve?

You are invited to participate in scenario workshop.

The purpose of the research is to understand last mile delivery difficulty within the Metropolitan Melbourne. You are requested to provide your view and opinion with regards to factors affecting last mile delivery. You will discuss how you perceive the difficulty created by the built environment and regulations (both planning and transport controls) on last mile delivery.

Participants will be requested to identify, rate and rank a range of built and regulatory environment attributes that hinder last mile delivery. Last mile delivery in this research means the final part of business to business delivery process. In addition, participants will be encouraged to discuss how the level of hindrance to last mile delivery can be classified from low to high impedance. Impedance in this research means level of impact or hindrance.

Specifically participants will:

i. Identify key indicators of the built environment and regulations that hinder last mile delivery. ii. Discuss and assess last mile delivery impedance indicators, and iii. Generate future scenario on impact of the indicators on last mile delivery.

1 Introduction

You are invited to take part in this research project, which is titled - Estimating last mile delivery impedance: A case of Melbourne. You have been invited because of your identification as part of the stakeholders in last mile delivery and your indication of interest to participate in the research scenario workshop.

Regulator - Your contact details were obtained from Municipal Association of Victoria (MAV);

235

Operator: Your contact details were obtained from the list of members of the Supply Chain and Logistics Industry Association in Metropolitan Melbourne.

End User (Community Groups) – Your contacts details were obtained from list of community groups in Metropolitan Melbourne.

This Participant Information Sheet/Consent Form tells you about the research project. It explains the processes involved with taking part. Knowing what is involved will help you decide if you want to take part in the research.

Please read this information carefully. Ask questions about anything that you don’t understand or want to know more about. Before deciding whether or not to take part, you might want to talk about it with a relative or friend.

Participation in this research is voluntary. If you don’t wish to take part, you don’t have to.

If you decide you want to take part in the research project, you will be asked to sign the consent section. By signing it you are telling us that you:

• Understand what you have read

• Consent to take part in the research project

You will be given a copy of this Participant Information and Consent Form to keep.

2 What is the purpose of this research?

 This research propose to measure and map impedance to last mile delivery on a transportation network using spatial indicators of transport and planning controls. Impedance to last mile delivery is defined as “the potential hindrance or obstruction to last mile delivery” as imposed by the built environment and regulatory environment (transportation and planning). This is in terms of constraints to movement of goods on a network and not in terms of time or monetary value. That is difficulty created by built and regulatory environments on last mile delivery.  Following the research, perceived last mile delivery impedance on attributes/indicators identified can be determined and map for efficient last mile delivery. The methodology used can be applied to estimate and determine last mile delivery in other cities.  The results of this research will be used by the researcher Kolawole Ewedairo to obtain a Doctor of Philosophy degree in Supply Chain/Logistics.  This research has been supported through an “Australian Government Research Training Program Scholarship”.  The research is not being coordinated outside the institution.

3 What does participation in this research involve?

 Before the commencement of the scenario workshop, participants will be invited to sign a consent form prior to involvement in the research.  No initial screening process is required.  Participants will be invited to discuss difficulty to last mile delivery created as a result of the built environment and the regulations (planning and transport controls) and how these difficulties can be mitigated.  The scenario workshop will consist of 15 participants.

236

 Half day participation with date to be determined after ethics approval and due consultation with participants.  Brainstorming and discussion  The activity will be audio recorded.  There will not be any use of an interpreter in the project.  The location of the scenario workshop will be building 80 at RMIT.

4 Other relevant information about the research project:

 You are taking part in a Research Scenario Workshop.  While the researcher will keep all the information relating to the workshop confidential (before or during workshop), the researcher cannot guarantee that other participants will do so.  We recommend that you do not share any personal or confidential information during the workshop.

5 Do I have to take part in this research project?

Participation in any research project is voluntary. If you do not wish to take part, you do not have to. If you decide to take part and later change your mind, you are free to withdraw from the project at any stage.

If you do decide to take part, you will be given this Participant Information and Consent Form to sign and you will be given a copy to keep.

Your decision whether to take part or not to take part, or to take part and then withdraw, will not affect your relationship with the researchers or with RMIT University.

If you take part in a scenario workshop, you are free to stop participating at any stage or to refuse to answer any questions. However, it will not be possible to withdraw your individual comments from our records once the workshop has started, as it is a group discussion.

6 What are the possible benefits of taking part?

There is no personal direct benefit from for participating in this scenario workshop from this research, however, you may appreciate contributing to knowledge.

There are various industry, community and societal benefit from participating in the scenario workshop. These benefits include facilitation of efficient last mile delivery, reduction in environmental impact of last mile delivery, reduction in traffic congestion resulting from last mile delivery.

7 What are the risks and disadvantages of taking part?

There is no risk envisaged from participating in the scenario workshop as no personal issues will be discussed.

8 What if I withdraw from this research project?

If you do consent to participate, you may withdraw at any time. If you decide to withdraw from the project, please notify a member of the research team.

237

You have the right to have any unprocessed data withdrawn and destroyed, providing it can be reliably identified.

9 What happens when the research project ends?

The outcome of the research will be provided to participants at the end of the project if requested. If participant have any question, the detailed contact information of the investigators will be provided to enable further discussion. How is the research project being conducted?

10 What will happen to information about me?

 The data collected will not be individually identifiable, re-identifiable (coded) or non- identifiable  Data will be stored for 5 years after publication of research findings within RMIT secured server after the completion of the project.  Only the researchers will have access to the worksheet and analysed data during the period of the research.  Data will only be used for the purposes described in the participant information sheet.  No identifiable particular of the participants will be recorded on the worksheet. Information on the worksheet will be responses from discussion and post-it paper used during the workshop.

By signing the consent form you consent to the research team collecting and using information from you for the research project. Any information obtained in connection with this research project that can identify you will remain confidential.

It is anticipated that the results of this research project will be published and/or presented in a variety of forums. In any publication and/or presentation, information will be provided in such a way that you cannot be identified, except with your express permission.

Any information that you provide can be disclosed only if (1) it is protect you or others from harm, (2) if specifically allowed by law, (3) you provide the researchers with written permission. Any information obtained for the purpose of this research project that can identify you will be treated as confidential and securely stored.

 Future use of data in 9.4 of the form.

The de-identifiable workshop worksheet and audio recording will be kept in secured RMIT server where it will be stored for a period of five years after the completion of the project.

The worksheet will contain recorded information and post-it paper from discussion during the scenario workshop. The worksheet is will not have any information about the participants.

11 Who is organising and funding the research? N/A

12 Who has reviewed the research project?

238

All research in Australia involving humans is reviewed by an independent group of people called a Human Research Ethics Committee (HREC). This research project has been approved by the RMIT University HREC.

This project will be carried out according to the National Statement on Ethical Conduct in Human Research (2007). This statement has been developed to protect the interests of people who agree to participate in human research studies.

13 Further information and who to contact

If you want any further information concerning this project, you can contact the researcher on 0417064099 or any of the following people:

Research contact person

Name Prof Prem Chhetri Position Senior Supervisor Telephone Email

14 Complaints

Should you have any concerns or questions about this research project, which you do not wish to discuss with the researchers listed in this document, then you may contact:

Reviewing HREC name RMIT University HREC Secretary Telephone Email Mailing address

239

Consent Form

Estimating last mile delivery impedance: A case Title of Melbourne.

Chief Investigator/Senior Prof. Prem Chhetri Supervisor Associate Investigators/Associates Prof. Jago Dodson and Prof. Paul Lee Supervisor Research Student Kolawole Ewedairo

Acknowledgement by Participant

I have read and understood the Participant Information Sheet.

I understand the purposes, procedures and risks of the research described in the project.

I have had an opportunity to ask questions and I am satisfied with the answers I have received.

I freely agree to participate in this research project as described and understand that I am free to withdraw at any time during the project without affecting my relationship with RMIT.

I understand that I will be given a signed copy of this document to keep.

I understand that audio recording of the workshop will be made for reference purpose.

Name of Participant (please Signatureprint) Date

Declaration by Researcher†

I have given a verbal explanation of the research project, its procedures and risks and I believe that the participant has understood that explanation.

Name of Researcher† (please print)

Signature Date

† An appropriately qualified member of the research team must provide the explanation of, and information concerning, the research project.

Note: All parties signing the consent section must date their own signature.

Note: It is recommended that you do not discuss any personal confidential issue during the workshop as the researcher cannot guarantee that other participant will keep such shared personal confidential information

240

Appendix B: Scenario Thinking Workshop Worksheet

Impact and Certainty worksheets

Scenario Building worksheet

241

Stakeholder Analysis worksheet

242

Appendix C: Location of Activity Centres

Councils Principal Activity Centres Major Activity Centres Location Banyule  Greensborough  Heidelberg  Ivanhoe Bayside  Brighton – Bay Street  Brighton – Church Street  Hampton  • Sandringham Boroondara  Camberwell Junction  Kew Junction Brimbank  Sunshine  Deer Park Central  Sydenham  Deer Park – Brimbank Central  St Albans

Cardinia  Pakenham Casey  Cranbourne  Endeavour Hills  Narre Warren – Fountain Gate Darebin  Preston – Northland  Northcote  Preston – High Street  Reservoir Frankston  Frankston  Karingal Glen Eira  Bentleigh  Carnegie  Caulfield  Elsternwick  Glenhuntly Greater  Dandenong  Chelsea Dandenong  Parkmore - Keysborough  Springvale Hobsons  Altona Bay  Altona North  Williamstown Hume  Broadmeadows  Gladstone Park  Roxburgh Park  Sunbury Kingston  Cheltenham – Southland  Cheltenham  Mentone  Moorabbin  Mordialloc Knox  Wantirna South – Knox  Bayswater City and Towerpoint  Boronia  Mountain Gate  Rowville – Stud Park

243

Manningham  Doncaster  Doncaster East – The Pines Maribyrnong  Footscray  Maribyrnong – Highpoint Maroondah  Ringwood  Croydon Melbourne  CBD  Carlton – Lygon Street Melton  Melton  Melton – Woodgrove and Coburns Road Monash  Glen Waverley  Clayton  Mount Waverley  Mount Waverley – Pinewood Centreway  Mulgrave – Waverley Gardens  Oakleigh  • Wheelers Hill Park Moonee  Airport West  Ascot Vale – Union Road Valley  Moonee Ponds  Niddrie – Keilor Road  North Essendon Moreland  Coburg  Brunswick  Glenroy Mornington  Hastings Peninsula  Mornington  Rosebud Nillumbik  Diamond Creek  Eltham Port Phillip  Balaclava  Port Melbourne – Bay Street  South Melbourne  • St Kilda Stonnington  Chadstone  Malvern/Armadale  Prahran/South Yarra  • Toorak Village Whitehorse  Box Hill  Burwood East – Kmart Plaza  Burwood East – Tally Ho Business Park  Forest Hill Chase  Nunawading Whittlesea  Epping  South Morang Wyndham  Werribee  Hoppers Crossing  Werribee Plaza Yarra  Fitzroy – Brunswick Street  Fitzroy – Smith Street  Richmond – Swan Street  Richmond – Bridge Road  Richmond – Victoria Street Yarra  Chirnside Park Ranges Lilydale

244

Appendix D: Planning Zones within Metropolitan Melbourne

Zones Nested Zones Residential Low Density Residential Zone (LDRZ) Mixed Use Zone (MUZ) Township Zone (TZ) Residential Growth Zone (RGZ) General Residential Zone (GRZ) General Residential Zone (GRZ) Industrial Zone Industrial 1 Zone (IN1Z) Industrial 2 Zone (IN2Z) Industrial 3 Zone (IN3Z) Commercial Commercial 1 Zone (C1Z or B1Z, B2Z, B5Z) Zone Commercial 2 Zone Rural Zone Rural Living Zone (RLZ) Green Wedge Zone (GWZ) Green Wedge A Zone (GWAZ) Rural Conservation Zone (RCZ) Farming Zone (FZ) Rural Activity Zone (RAZ) Public Land Public Use Zone (PUZ1-7) Zone Public Park and Recreation Zone (PPRZ) Public Conservation and Resource Zone (PCRZ) Road Zone (RDZ1&2) Special Special Use Zone (SUZ) Purpose Zone Comprehensive Development Zone (CDZ) Urban Floodway Zone (UFZ) Capital City Zone (CCZ) Docklands Zone (DZ) Priority Development Zone Urban Growth Zone (UGZ) Activity Centre Zone (ACZ) Port Zone (PZ) Source: Victoria Planning Provision

245