A Critical Review on Population Synthesis for Activity- and Agent-Based Transportation Models

Abstract

Traditional four-step transportation planning models fail to capture novel transportation modes such as car/ridesharing. Hence, agent-based models are replacing those traditional models for their scalability, robustness, and capability of simulating nontraditional transportation modes. A crucial step in developing agent-based models is the definition of agents, e.g., household and persons. While model developers wish to capture typical workday travel patterns of the entire study population of travelers, such detailed data are unavailable due to privacy concerns and technical and financial feasibility issues. Hence, modelers opt for population syntheses based on travel diary surveys, land use data, and census data. The most prominent techniques are iterative proportional fitting (IPF), iterative proportional updating (IPU), combinatorial optimization, Markov-based and fitness-based syntheses, and other emerging approaches. Yet, at present, there is no clear guideline on using any of the available techniques. To bridge this gap, this chapter presents a comprehensive synthesis of practice and documents available successful studies.