EXPERIMENTAL INVESTIGATIONS OF A PERCEPTION UPDATING MODEL ON A TRAFFIC NETWORK WITH INFORMATION PROVISION

In this paper, a simulation framework is presented for studying day- to-day dynamics in traffic networks with advanced information systems. Each driver has their own travel time distributions which describe their own time varying perceptions of the network performance. Perceptions are updated through a Bayesian updating model. Drivers' route and departure time choices are based on their perceived travel times in the network. Choice probabilities are estimated through a nested logit model. DYNASMART, a traffic simulation model, is used to investigate the perception updating model within a day-to-day framework.