A formulation and solution algorithm for multiple user classes dynamic traffic assignment

Dynamic traffic assignment (DTA) is the base of real-time traffic control and management. However, the present work on multi-class analytical DTA model assumes that the link travel time function is identical for all travelers. Again the interactions among multiple classes are neglected. Moreover the different route choice rules to different users are hardly considered. Therefore, a multi-class analytical DTA model is proposed in this paper. The user classes typically refer to different classes whose route choice rules are different, such as dynamic system optimization, dynamic user equilibrium, and stochastic dynamic user equilibrium. According to travel time characteristic of each user class, the multidimensional travel time function based on the extended BPR function is built which can satisfy FIFO requirement and express the asymmetric interactions among user classes. At the same time, the interdependences among user classes are modeled based on bi-level programming, and the solution algorithm on bi-level model is designed using iterative optimization algorithm. Finally, the model is applied to the experimental network.