Study On The Model Of Urban Transportation Integrated System Based On Multi-agent And Game Theory

Urban transportation which plays a significant role in the normal function of a city is one of important infrastructural facilities. The sustained development of urban economy has been affected by transportation problems, which have been drawn much attention. AH traditional measures to solve the urban transportation problems have failed to meet the existing requirements of the urban transport development. With taking into consideration the complexity of urban transportation, the limitations of traditional traffic control methods and the application of artificial intelligence technology, the optimization technology of urban traffic control system along with multi-agent technology, genetic algorithm and game theory has been discussed in this paper.Firstly, the development of cooperative control methods of domestic and foreign urban transportation system have been analyzed, especially in the feature, the composition model of agent and cooperative methods of multi-agent. With the characteristics of multi-agent, the possibility and superiority of multi-agent technology used in the cooperative control of urban transportation system have been pointed out.Secondly, an analysis of the existing traffic mode, control structures and traffic control system based on multi-agent has been made. With the deficiency of the existing traffic control model based on multi-agent technology, the control authority of agent in the traffic cooperative control system and the functional structure of intersection agent have been discussed. The multi-agent traffic coordination control system structure based on intersection agent has been constructed. Next, the intersection agent optimization model on the basis of genetic algorithms of real-code with the average vehicle delay of the intersection as objective function has been established under the frame of the traffic cooperative control system structure based on multi-agent. "The optimized strategy was better than the existing actual control strategy" has been showed in the paper on the condition of utilizing the model to optimize the signal phasing and timing. Finally, via game theory, the cooperative mechanism of the simple network composed by double intersection agent and the game process of intersections have been explored. Then, the game process of intersections has been established under analyzing the respective problems of these two intersection agents. The control strategy of intersection has been optimized according to the actual intersection situation. At the same time, the feasibility and availability of the cooperative mechanism has been showed based on the comparison between the optimized function index and the existing control made in the paper.