In order to solve the key problems of urban traffic in China, the intelligent means and approaches, combined with artificial intelligent with tradition control ones, are adopted in the paper. Some hypotheses are elicited on the basis of the characteristics of Chinese urban traffic control system structure and functional requirement, a frame structure of urban traffic intelligent control system is stated based on multi-agents cooperation. Some better design ways and means for urban traffic real-time, rational and reasonable control system is established according to the frame structure, the system design of urban cross control and area control based on multi-agents is implemented. Simulation results show that urban traffic integration control with multi-agent excels timing control in latency time. In this way, successful realization of goal for urban traffic intelligent control is insured so as to increase the capacity of urban road network and to improve our urban traffic control and management modes. Therefore quality of urban surroundings will be enhanced. And the comfortable and delightful traffic surroundings will be built.
Urban Traffic Control Using Adjusted Reinforcement Learning in a Multi-agent System
