Research On Day-To-Day Dynamic Traffic Assignment Model For Muti-Case

With fast development of urbanization and rapid increase of vehicle ownership, the carrying capacity of urban roads may reach the limit, and road congestion is increasing day by day. By researching the urban traffic network, we can analyze traveler behavioral and psychological characteristics. We can also understand the formation mechanism of congestion and study the change tendency of traffic flow by building proper dynamic traffic assignment evolutionary models, which provide the effective basis theories for traffic control and management.Around the problem about the dynamic traffic assignment, this paper focused on three aspects as follows:1. A dynamic assignment model based on the bottle-neck theory is proposed. It can formulate the evolutionary process that travelers adjust their departure time from day to day. Then, we give the processes of solution and validation, and use numerical experiments to show the effectiveness of the model. The results show the trajectory of the model can converge to a fixed point, and the stationary sate of the model is corresponding to the deterministic user equilibrium state. Moreover, the model can depict the distribution of travelers in each interval.2. Considering the traffic network structure may be changed due to the exogenous factors, we propose a dynamic assignment model to describe the adjustment process of travelers’route choice. We analyze several properties of the model and use numerical experiments to illustrate the properties of the model. In addition, the model can also analyze the degree of importance of links, which provides the basis for redesign and management of the traffic network.3. Considering travelers are boundedly rational, we propose a discrete dynamical system model. Its stationary state is equivalent to the Logit stochastic user equilibrium. We also analyze several properties of the model, including the equivalence between the stationary state and user equilibrium and the uniqueness and convergence of its stationary state. A method of validating the parameter of this model is introduced. The model is given in general form and hence we introduce a special case of the model. Finally, numerical experiments are conducted to demonstrate the application and properties of the model.