Transferability of an Activity-Based Travel Demand Model and Heterogeneous Preferences in Mode Choice Behavior: A Study for Ho Chi Minh City, Vietnam

Abstract

The rapid development of economic activities in big cities of the developing countries in Southeast Asian has drastically increased the travel demand. Effective travel demand management (TDM) measures and sensitive travel demand models to evaluate TDM measures as the activity-based travel demand models (ABM) are important for these cities to deal with the chaos of the traffic with congestion and accidents due to the mixed of cars and motorcycle. However, the development of an ABM from scratch requires significant resources in terms of time, money and skills. Therefore, a promising alternative is to use an already existing well-developed ABM system. Studies on the transferability of an activity-based model in the settings of developing countries are nevertheless missing in the literature. Therefore, the main goals of this study are to investigate the transferability of a well-developed ABM system and modify the transferred model to be workable in the context of the motorcycle-based cities, based on availability of local data resources. The transferred context is Ho Chi Minh city, Vietnam (HCMC), a typical motorcycle-based city. Among many operating ABMs, FEATHERS (Forecasting Evolutionary Activity – Travel of Household and their Environmental RepercurssionS) originally developed for Flanders, Belgium (Bellemans et al., 2010) is chosen as the base model because it is a fully operating model and opened for research purpose. However, adopting an ABM does not mean that all the modeling and policy evaluation problems have been solved. Because, an ABM is a “computational framework for modeling behavior, rather than a model per se” (Miller, 2019, pp 129). An ABM should be based on desired behavioral realism reflecting the local context in order to explain policy impacts. Heterogeneity is also among ongoing challenges to the improvement of ABM modeling systems (Miller, 2019), although it has been somehow handled in ABM systems due to the individual oriented modeling nature of ABMs. In the context of the motorcycle-based societies where are occurring the transition from motorcycle to automobile usage (Nishitateno & Burke, 2014), the heterogeneity of the road users also significantly contributes to the diversity in responses to travel demand management policies. The exploration of this distinct mechanism is another important issue that requests for further research. To address these issues, two main objectives of this thesis are: 1) to develop a modeling framework of the first ABM for HCMC based on FEATHERS utilizing available data resources and 2) to explore the potential source of road users’ heterogeneous tastes toward road pricing policies so as to support further improvements of an ABM for a motorcycle-based cities as HCMC. Regarding the first objective, this thesis presents the effort to transfer FEATHERS to HCMC and explore the issues that require further development of travel demand modeling tools for the developing countries. The first part of the research investigated the systematic (dis-)similarities of the transportation behavior between Flanders (for which FEATHERS system has already developed) and HCMC. This provides an extent of the transferability and data requirements when implementing an ABM in a distinctive cultural setting such as HCMC. The transferability was examined through the comparison of important activity – travel behavioral indicators at three modeling levels (daily pattern, tour and trip). The results confirmed that the location and transport mode choices are the least transferable models. Further examination also found that the modeling structure which reflects the interrelationship among travel choices is also not transferrable. In the next step, behavioral models and their modeling structure were estimated and embedded in FEATHERS – HCMC, the FEATHERS version for HCMC. Several motorcycle travel demand management measures have been tested to validate the capability of the model in analyzing policy impacts. To address the second objective, an investigation of the behavioral responses to pricing policies from private car users and motorcycle users was implemented. A stated-preference experiment on transport mode choice with varied pricings was conducted in HCMC. The results revealed the existence of a transitional group of motorcycle users who will switch to being car users. This finding suggested the high diversity among the road users that cannot be ignored by policy makers in the design of future pricing policies as well as the need to improve the capability of travel demand models. In summary, the original contributions of this study are as follows: 1) This is the first ABM modeling framework that developed for Ho Chi Minh city of Vietnam. It serves as a prototype framework for the developing countries with limitation in data and expertise resources; 2) The study is also the first to identify a group of transitional motorcycle users and its impacts on policy implementation during the transition to automobiles of the MBCs. This finding not only contributes more insights to transport planners in policy design process but also suggests for the careful consideration in an ABM modeling system for similar contexts.