Modeling Routing Behavior Learning Process for Vacant Taxis in a Congested Urban Traffic Network

In this paper, we present a modeling framework and approach to capture vacant taxi drivers' route choice behavior learning process and simulate their changes of routing decisions over time due to updated experiences of the traffic and passenger's information. Efforts to unveil their behavioral learning process were rather limited, although some researchers focused on the modeling of routing behavior. We focused on the street-hailing of vacant taxi drivers, who selected a route to minimize the search time for picking-up a waiting customer along the road, which was determined by the traffic information and customer arrival rate. At the end of each learning cycle, or "learning day," taxi drivers updated their knowledge on the traffic and passengers based on their newly gained experience, and made corresponding changes to their route choice at the next learning day until an optimal route had been found. Both analytical and numerical analysis were conducted on the Taipei traffic simulation network. The case study results showed that the proposed model was able to reasonably capture taxi drivers' changes of route choice.