Study On Equilibrium Model And Algorithm For Ubran Degradable Traffic Network

The uncertainty of urban traffic network is often neglected by the traffic re-searchers nowadays. How to accurately provide the information of the shortest path for travelers under the current traffic condition becomes a pressing problem for the intelligent traffic researchers. The main goal of the research on degrad-able traffic network is providing real-time traffic information for travelers. Thus, the most suitable paths for travelers are chosen to avoid the congested roads and unnecessary loss. In this thesis, the traffic assignment problem with en-route diversion is studied in the degradable traffic network. Meanwhile, a new algorithm is proposed for solving the equilibrium assignment models.Firstly, the loss amount of the network capacity caused by the degraded roads is analyzed. Then, the path selection and the en-route diversion are stud-ied when the traffic information changes in the degradable traffic network. The change of the traffic flow is obtained based on the above analysis. The stochastic user equilibrium model with elastic demand is established in the degradable traffic network. The equivalent of the model to the SUE condition is proved. A simple example is given to illustrate the availability of the model.In order to meet the need of intelligent traffic research, a new algorithm for solving the equilibrium assignment models is given based on the advantage of the path algorithm providing the shortest path information. A modified projec-tion gradient method with a new step length is combined with the path genera-tion method in the algorithm. The feasibility and efficiency of the algorithm are showed by numerical examples in several simply network. The traffic flow is clearly described by the path information from the new algorithm, which can lay the foundation for the further research on path choice.