A Study Of Urban Traffic Optimization Based On Stackelberg Game Model

As economy roars, urban scale extends and urban population surges, urban roads are facing anincreasing pressure. Traffic congestion has already become a Gordian knot that metropolis has todeal with everyday. The smoothness and the efficiency of traffic network are the key factors ofurban traffic network.Since smart traffic information system develops rapidly, route guidance has been an importantmeans to optimize the traffic so that the traffic manager could provide a rational guidance for thetraveler by issuing the traffic information. However, in real life, the guidance of road network failsto reach the ideal condition between the manager and the traveler.Taking the traffic optimization as the cut-in point, the thesis discusses the interest contradictionbetween the manager and the traveler. The manager issues the traffic information based on SO,while the traveler, as individuals with subjective initiative, prefer to the optimization of their own.On the basis of this contradiction, a Stackelberg game model of the manager and the traveler hasbeen build in this thesis. In the game model, the manager is pursuing the minimum ofcomprehensive travel cost of traffic network, which includes the time cost and the fuel cost; thetraveler is pursuing self-profit maximization, which is evaluated through utility theory. The resultwould be solved by an enlightening algorithm. Meanwhile, an innovation for improvement antcolony algorithm is proposed, which combined the route choice probability of the traveler andconsidered the live-traffic report. This improvement would be used as the solving algorithm of themanger’s optimization. Ultimately, through the mutual game between the manager and the traveler,the efficiency of the traffic network would be improved and the distribution of the cars balanced.In the end, a simple traffic network is designed to carry on simulation test by MATLAB. Thereare containing three contrastive simulation tests: shortest path trip mode without guidance, best timetrip mode without guidance and comprehensive trip mode with guidance. By analyzing the results,it draws a conclusion that the traffic can be smoother and more balanced, and comprehensive costwould be declined under the guidance of the model built in this thesis instead of the model that onlybased on the experience of the traveler self.