Temporal Transferability Assessments of Vehicle Ownership Models and Trip Generation Models for Boston Metropolitan Area by Yafei Han B.E., Urban Planning and B.A., Economics, Peking University, 2011 Submitted to the Department of Civil and Environmental Engineering and the Department of Urban Studies and Planning in partial fulfillment of the requirements for the degrees of Master of Science in Transportation and Master in City Planning at the Massachusetts Institute of Technology June 2015 © 2015 Massachusetts Institute of Technology. All rights reserved Author: ............................................................................................................................................... Department of Civil and Environmental Engineering Department of Urban Studies and Planning May 21, 2015 Certified by......................................................................................................................................... P. Christopher Zegras Associate Professor, Department of Urban Studies and Planning Thesis Supervisor Certified by......................................................................................................................................... Mikel Murga Research Associate, Department of Civil and Environmental Engineering Thesis Supervisor Accepted by ....................................................................................................................................... Dennis Frenchman Professor, Chair MCP Committee, Department of Urban Studies and Planning Accepted by ...................................................................................................................................... Heidi Nepf Donald and Martha Harleman Professor of Civil and Environmental Engineering Chair, Graduate Program Committee Temporal Transferability Assessments of Vehicle Ownership Models and Trip Generation Models for Boston Metropolitan Area by Yafei Han Submitted to the Department of Civil and Environmental Engineering and the Department of Urban Studies and Planning on May 21, 2015 in partial fulfillment of the requirements of the degrees of Master of Science in Transportation and Master in City Planning at the Massachusetts Institute of Technology Abstract In the last few decades, travel demand models have undergone tremendous development and, today, are routinely used to support planning and policy decisions. But uncertainty in forecasting with such models is often overlooked, and its impact on forecast accuracy is rarely evaluated. My thesis attempts to understand behavior uncertainty and model uncertainty in travel demand modeling. In particular, I assess the temporal transferability of vehicle ownership models and trip generation models for the Boston metropolitan area from 1990 to 2010. Through statistical tests, I find significantly changed preferences in household vehicle ownership choice and trip production. For vehicle ownership choice, the effects of most socio-economic and demographic factors, and regional location factor have evolved; while the effects of local built environment factors and transit access are stable. Trip rates have changed over time, with decreased home-based work, home-based shopping, home-based bank and personal business, home-based social, home-based eating and non-home-based work trips; and increased home-based recreational and home-based work-related trips. 2 The prediction tests suggest that failing to consider preference changes causes significant bias in forecasts. The transferred vehicle ownership model of 1991 under-predicts 0- vehicle households by 42.5%, and over-predicts 2-vehicle households by 14.8% in 2010. The transferred trip rates from 1991 overestimate total trips in 2010 by 7% to 9%. Home- based work-related, home-based pick-up and drop-off, and home-based recreational trips are significantly under-predicted by 34%, 12% and 27%, respectively; while home-based work, home-based shopping, home-based social, and non-home-based work trips are significantly over-predicted by 9%, 20%, 31%, and 69%, respectively. Different model specifications have shown a modest range of variability in prediction outcomes, suggesting model specification uncertainty has less influence than behavior uncertainty on forecasts. In vehicle ownership modeling, children, seniors, and local built environment variables improve the prediction accuracy for 0-vehicle group. But all model specifications cannot distinguish well between 0- and 1-vehicle households, and between 2- and 3-vehicle households. Household characterization affects the prediction accuracy for certain trip purposes. Including more detailed household information may lead to worse forecasts because of large sampling variance. Future work is suggested to incorporate behavior uncertainty in forecasts, explore uncertainty in model structure, and evaluate the practical implications of the lack of model transferability. Thesis Supervisor: P. Christopher Zegras Title: Associate Professor of Urban Studies and Planning Thesis Supervisor: Mikel Murga Title: Lecturer and Research Associate, Civil and Environmental Engineering 3 Acknowledgements This work and my time at MIT would not have been such a valuable experience without the guidance, support, wisdom and love of many people. My thesis would have been impossible without the guidance and insights of my advisors. My academic and thesis advisor Chris Zegras opened up this fascinating research field of uncertainty to me two years ago. He inspired me with many interesting questions, and provided countless insights and detailed instructions. He offered critical advice for my education at MIT, and provided abundant resources and connections with other researchers and institutions, all of which have nourished this thesis. I greatly appreciate and admire his generosity, thoughtfulness and patience in his mentorship. Mikel Murga enthusiastically shared his expertise and rich experiences in transport modeling. His excellent teaching of four-step modeling not only equipped me with the modeling techniques, but also led me to think through the model assumptions and sources of uncertainty. He generously offered the Cube four-step model he created for Boston metropolitan area, which was of great use in this thesis. His emphasis on empirical realism also helped shape this thesis. I give thanks to Prof. Ben-Akiva for his teaching of demand modeling, and his advice on preference stability test in my thesis. I am thankful to Fred Salvucci, John Salvucci, and Jinhua Zhao for their constructive feedbacks to my presentation in Transit Lab. I am very grateful to the Masdar Institute for funding this research. Thanks to Caliper and Citilabs for offering me academic license for using their software. I am very thankful to my research teammate Michael Dowd for his countless help with software issues (e.g. Cube, TransCAD, python etc.). He has always been available and selflessly sharing his time and energy. I greatly respect his generosity and responsibility. I am also thankful to Victor Rocco, Pablo Posada Marino, Shenhao Wang and Menghan Li for their advice and help with this thesis. I am very grateful to the family of God in Boston. Especially I thank my current leader Irene unni, and former leaders – Jeansue unni and Insue unni for their love, guidance and 4 prayers. I thank my fellow sisters at Harvard-MIT bible study group and in sisters’ house, for their presence in my life, and all the time we spent together. I thank them for sharing my burdens, cooking and giving rides to me during my difficult times. I am deeply indebted to my parents for supporting me to come to Boston. Without their endless love poured out for me, I could not have gone so far. Finally, I praise and thank God for His presence in my life, and for His truth and grace. 5 Contents Chapter 1 Introduction ...................................................................................... 9 Chapter 2 Literature review ............................................................................ 11 2.1 Travel demand forecast inaccuracy ................................................................... 11 2.2 Sources of uncertainty ...................................................................................... 16 2.2.1 Exogenous input uncertainty ....................................................................... 16 2.2.2 Behavior uncertainty ................................................................................... 19 2.2.3 Model uncertainty ....................................................................................... 19 2.3 Uncertainty analysis approaches ....................................................................... 20 2.3.1 Uncertainty propagation .............................................................................. 20 2.3.2 Model transferability assessment ................................................................ 25 2.4 Empirical findings: temporal transferability of vehicle ownership models and trip generation models ..................................................................................................... 30 2.4.1 Temporal transferability of vehicle ownership model ................................... 30 2.4.2 Temporal transferability of trip generation model ........................................ 32 2.5 Summary .......................................................................................................... 35 Chapter 3 Methods .........................................................................................