The impact of bus priority policies on peak commuters behavior: An agent-based modelling perspective

Abstract

By using micro-simulation method and the BushCMosteller reinforcement learning model, this paper modeled the behavior of urban commuters departure time choice on a many-to-one transit system during the morning peak-period. Three kinds of typical urban public transport priority policies were studied. Result shows that if we can choose the right time for free public transportation, the pre-peak-free policy will have certain effects on staggering the commuting peak by influencing commuters decision-making on departure-time. As for the bus-accelerating policy, it can lower commuters cost, but it is likely to cause more congested volume and add more pressure on the public transit system. The departure-frequency increasing policy can partially alleviate the peak congestion problem, but cannot fundamentally eliminate the congestion, instead, it may increase the operating costs. This research is helpful in acquiring a better understanding of commuters departure time choice and commuting equilibrium during the peak-period. The research approaches also provide an effective way to explore the formation and evolution of complicated traffic phenomena.