Simulation And Research Of Urban Dynamic Traffic Assignment Model Considering Crowding Effects

Dynamic traffic assignment is an important research subject of the current transportation economics and model analysis of traffic network, which possesses efficient time varying variable demand of traffic assignment in traffic network, so it has the primary guidance value for vehicle optimal route, traffic guidance and vehicle travel time. Although the traditional algorithms, such as all-or-nothing assignment method, incremental assignment method,method of successive averages, Frank-Wolfe method, can distribute traffic to various sections,these methods can not dynamic simulate route choice behavior of vehicles in the road network,which leads to traffic assignment different from the reality. However, ant colony algorithm has strong flexibility and robustness in a dynamic environment. This paper will use the ant colony algorithm to solve dynamic traffic optimal control model. Route guidance and traffic assignment of vehicles are achieved by designing algorithm. It also analyzed route choice behavior of vehicles and traffic changes of each section in road network combined with SUMO simulation when occurs different traffic conditions. The ant colony system has strong flexibility and robustness in a dynamic environment, which is used to solve the dynamic traffic optimal control model in this paper. Compared with the traditional method, it can better choose the right path for induction and realize the flow distribution. The main work of this paper is as follows:1. Using method of successive averages and ant colony algorithm to achieve dynamic traffic system optimal model, traffic of each section is obtained respectively by the two methods. Through analyzing and comparing result of flow distribution, it concluded that ant colony algorithm can more evenly assign traffic to the road network and more advantageous for planning and designing of traffic management compared with the method of successive averages.2. Using Frank-Wolfe and ant colony algorithm to achieve dynamic traffic user optimal model, traffic of each section is obtained respectively by the two methods. Through analyzing and comparing result of flow distribution, it concluded that ant colony algorithm can more evenly assign traffic to the road network and more advantageous for travel of users comparedwith the Frank-Wolfe method.3. In order to verify that ant colony algorithm is effective in the dynamic traffic system optimal model and dynamic traffic user optimal model, we use SUMO to simulate traffic in road network. Through analyzing simulation result of traffic, it concluded that ant colony algorithm is applicable to solve the dynamic traffic control optimal model. And analyzing route choice of vehicles is how to impact result of traffic assignment, when occurs accidents on the road. As well as analyzing traffic changes when traffic suddenly increase or reduce on each section.