Research On Morning Peak Commuters Based On Multi-Agent Simulation

The agglomeration phenomenon of urban residents is more and more serious, especially during morning and evening peak hours. Traffic congestion problem resulted from the phenomenon has seriously affected residents' daily life and brought issues such as air pollution and lack of parking. As an important component of traffic demand management, research on commuters' departure time can help us to fully understand the influence factors and behavioral mechanisms of commuters' departure time choice, and it is a prerequisite for designing and evaluating traffic policy.By combining Multi-Agent simulation technology and the Bush–Mosteller reinforcement learning model, the simulation model of commuters' departure time choice in urban morning peak hours was established. By comparing with results of the classic analytical model under the same conditions, the correctness of the model was verified, and the formation process of departure time equilibrium was reproduced. Two kinds of simulation experiments were carried out from aspects of the commuter and the environmental policies. In the experiment about factors of commuters themselves, commuters' heterogeneity and memory limitations were mainly considered. And in the experiment about factors of environmental policies, effects of "free before peak hours" policy and "bus priority" policy on departure time equilibrium were investigated emphatically.The research is helpful in acquiring a better understanding of commuter's departure time choice and commuting equilibrium of the peak-period, the approach also provides an effective way to explore the formation and evolution of complicated traffic phenomena.