Network Traffic Flow Evolution Model Considering OD Demand Mutation

Abstract

Considering such a fact that the traffic demand freshly entering a network does not have perfect information about traffic condition and then might choose travel routes randomly, we propose an improved network traffic flow evolution model in this article. The model's properties and fixed points are investigated. The fixed points of the model, which are different from traditional user equilibrium state or part user equilibrium state, can better formulate the result of traffic flow assignment in realization. Simulation results obtained from a grid network show that the model can be used to approximately model the process of realizing user equilibrium state, the network flow can indeed evolve to one or more fixed points, and the initial values of all available path flows can be zero.