Hybrid Agent based Simulation with Adaptive Learning of Travel Mode Choices for University Commuters (WIP)

摘要

This paper presents a methodology for developing a hybrid agent-based micro-simulation model to capture the impacts of commuter travel mode choices on a University campus transport network. The proposed methodology involves: (ⅰ) developing realistic population of commuter agents (students and staff); (ⅱ) assigning activity lists and travel mode choices to agents using machine learning method; and, (ⅲ) traffic micro-simulation of the study area transport network. This furthers the understanding of current transport modal distributions, factors affecting the travel mode choice decisions, and, network performance through a number of hypothetical travel scenarios.