A game theory based traffic assignment using queueing networks

Traffic assignment models can be classified according to the behavioral assumption governing route choice: the User Equilibrium (UE) and System Optimum (SO) traffic assignment. The first model refers to the “user-optimal” model, that is the users´ least cost model, while the latter is based on overall cost minimization. The paradox of benefit between UE and SO makes it impractical. In addition, the dynamic and stochastic characteristic of traffic flow in the congested traffic network makes it difficult to be modeled by the deterministic method. It is necessary to make queueing analysis in the congested network. This paper presents a game theory based traffic assignment using queueing networks, which can balance between UE and SO and give more feasible traffic assignment. A comparative experiment is conducted among SO, UE and noncooperative models to validate the model.