Understanding and Optimising Carsharing Systems Sisi Jian
Total Page:16
File Type:pdf, Size:1020Kb
Understanding and Optimising Carsharing Systems Sisi Jian A thesis presented in fulfilment of the requirements for the degree of Doctor of Philosophy School of Civil and Environmental Engineering The University of New South Wales August 2017 i Originality Statement ii COPYRI GHT STATEMENT 'I h reby � grant the University of New South Wales or its agents the right to rch,ve an o � � t ma e avail b e or art in the _ . _ � a le my th sis or dissertation in whole � n,v ,ty b �� lr ranes 1n all forms of media, now or here after known, subJect to the rov,s,o � ns of the Copyright Act 1968. I retain all proprietary rights, such as patent rights . I also retain the right to use in future works (such as articles or books) all or parto f this thesis or dissertation. 1 lso � authorise University Microfilms to use the 350 word abstract of my thesis in Dis e s rtation Abstract International (this is applicable to doctoral theses only). I e hav either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been g ranted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.' Signed ············ ··················· · ········ ··· Date ............ ;?. 1 .. 1 .. P.� --· .l. .. t.?t...... ......................... .. AUTHENTICITY STATEMENT 'I certify that the Library deposit digital copy is a d!rect equivalent of the final officially approved version of my thesis. No emendat,on of content has occurred _ and if there are any minor variations in formatting, they are the result of the conversion to digital format.' Signed ' , ................................ ···············�················ 1 Date ............-2.q ! .. I········· o� .f.. .l .......................................... Abstract Carsharing, as an alternative to private vehicle ownership, has spread worldwide in recent years due to its potential of reducing congestion, improving auto utilisation rates and limiting the environmental impact of emissions releases. Along with its growth, the flexibility of carsharing systems also brings complex problems to the operators. One dominant challenge in carsharing systems is to ensure the supply of vehicles can meet the demand of users in a cost-effective manner. This requires accurately predicting users' demand and optimally relocating vehicles in response to demand variations. The two principal areas of this thesis are methods to estimate demand and optimally relocate fleet. From the demand side, this study models users’ vehicle selection and utilisation patterns. Focusing on vehicle selection behaviour, a spatial hazard-based model (SHBM) is proposed to investigate the impacts of users’ socio-demographic attributes and fleet characteristics on their choice set formation behaviour in selecting vehicles. The modelling is achieved by regarding “distance to carsharing vehicle” as a random variable analogous to the duration in conventional hazard-based models. Data collected from the Australian carsharing company GoGet are utilised to calibrate the models. The accelerated failure time model with a log-logistic distribution is found to provide the best fit. Upon making a vehicle selection, users then decide the amount of consumption to allocate to each selected vehicle type. This process involves making multiple discrete choices of continuous amounts and is modelled by the multiple discrete-continuous extreme value (MDCEV) modelling framework. Three MDCEV models considering travel time, mileage, and monetary expenditure as the continuous i consumption constraints are developed to estimate the impacts of a set of socio- demographic attributes on user’s vehicle choice and capture the satiation effect with increasing the consumption for each vehicle type. An efficient simulation procedure is applied to evaluate the performance of the three MDCEV models. The results indicate travel time, mileage and expenditure affect users’ vehicle usage pattern in the same way. The findings from these two demand models can be referred to by the operators when determining the most efficient allocation of resources within carsharing systems. From the operation side, the research develops and solves novel models for the vehicle stock imbalance problem in one-way carsharing systems. Previous studies have proposed relocation methods to handle it, but the interdependence between demand and supply has never been considered. The thesis proposes two relocation models to link demand and supply. Both incorporate a discrete choice model (DCM) in an integer linear programming (ILP) model to account for the interaction. The difference between them lies in the DCMs. In the first model, the DCM does not assume users’ demand to be elastic to vehicle availability. The ILP model solves optimal relocation decisions and updates vehicle availability for each station; the DCM then coupled with the updated vehicle availability changes users’ trip demand reciprocally. Built on the first model, the second model extends the DCM by including vehicle availability as a parameter directly affecting demand. In this new framework, demand and supply are linked by vehicle availability: it is the output of the ILP model and at the same time the input of the DCM. The nonlinearity of the DCM is further linearised through a linearisation approach. Both models are tested in the GoGet network. The results reveal if there is a strong interdependence between demand and supply, the supply has a critical impact on system profit. ii The core contribution of this thesis is to take the first attempt to understand and optimise carsharing systems considering the interdependency of demand and supply comprehensively. iii Acknowledgements I will never forget the lucky Skype call I received four years ago when I was studying at Singapore. Yes, that’s the interview call from my supervisor Dr. Vinayak Dixit, and that Skype call led me to the world of research and changed my life entirely. I am so fortune to have Vinayak to be my supervisor. I can hardly find proper words to express how much it means to me for his academic suggestions, career advices, constructive criticisms, and giving me precious opportunities to see the world. Also, I would like to express my most sincere thanks to my supervisor Prof. Travis Waller, for providing so many invaluable suggestions and inspiring me to pursue the happiness of research. PhD study is a tough process, but I am so grateful to have the most lovely colleagues in our group sharing the gains and frustrations during the entire process. Specially, I would like to thank Dr. Taha Hossein Rashidi and Dr. David Rey for their fantastic advices on my researches. I would like to thank Dr. Kasun Wijayaratna for all those long talks, and his kind advices not only on researches, but also on my life in Australia. I would also like to thank Ms Maria Lee and Ms Sylvia Brohl for providing incredible assistance in dealing with so many paper works. I wish to also thank Dr. Zhitao Xiong, Xun Li, Chenyang Li, Xiang Zhang, and Tao Wen for wines and hot pots. Good wines and delicious food are the best medicines to relieve the tensions from work. Furthermore, I would like to thank Mr. Bruce Jeffreys and Ms. Rachel Moore from GoGet for providing the data to support my research, and the Australian Research Council for their support under Linkage Grant. Finally, I am so lucky to have the greatest husband who always supports me and tolerates me with his love, understanding, and patience. Also, I would like to thank iv my parents for their selfless love and understanding my decision of studying abroad. I wish to also thank my two cats for accompanying me all the time and bringing me so much happiness. My dearest family always give me the light when I am in shadow. v List of Relevant Publications The following provides a list of the journal and conference publications that have contributed towards the development of the thesis. Peer-reviewed journal publications 1. Jian, S., Rashidi, T.H., Wijayaratna, K.P. and Dixit, V.V., 2016. A Spatial Hazard-Based analysis for modelling vehicle selection in station-based carsharing systems. Transportation Research Part C: Emerging Technologies, 72, pp.130-142. 2. Jian, S., Rey, D., Dixit, V., 2016. Dynamic Optimal Vehicle Relocation in Carsharing Systems Accepted for publication in the 2016 Transportation Research Record: Journal of the Transportation Research Board. 3. Jian, S., Rashidi, T.H., and Dixit, V., 2017. An analysis of carsharing vehicle choice and utilization patterns using multiple discrete-continuous extreme value (MDCEV) models. Transportation Research Part A: Policy and Practice, 103, pp.362-376. Papers submitted and under review 4. Jian, S., Rey, D., and Dixit, V. An Integrated Supply-Demand Approach to Solving Optimal Relocations in Carsharing Systems. In review with Networks and Spatial Economics. Peer-reviewed conference papers 5. Jian, S., Rey, D., and Dixit, V., 2016. Dynamic Optimal Vehicle Relocation in Carsharing Systems. In proceedings of the 95th Transportation Research Board Annual Meeting, Washington, D.C., 10 - 14 January 2016. 6. Jian, S., Rashidi, T.H., Wijayaratna, K. P., and Dixit, V., 2015. Hazard-based modelling of vehicle selection in carsharing systems. In proceedings of the 94th Transportation Research Board Annual Meeting, Washington, D.C., 11 - 15 January 2015 7. Dixit, V., Trieu, J., Jian, S., and Li, X., 2014. Value of travel time savings for carsharing users in Sydney, presented at Australian