Research Of Urban Road Transportation Congestion Coordination Control Based On Multi-Agent Technology

With the rapid development of national economy and the pace of urbanization speeding up, traffic’s influence on the urban economic development and the improvement of people’s living level is increasingly important. As the urban road construction is difficult to keep up with the speed of traffic development, the rapid increase of vehicle and the growth of the road presents a large contradiction, the severity of the traffic congestion has also increased sharply. The main way to solve the congestion of traffic is:traffic management facilities as a starting point, scientific organization scientific control of road traffic flow, make its maximum orderly flow, the traffic potential of the road network traffic plays to the largest. Given the limitations of traditional traffic control mode and the complexity of traffic scale, people using a higher level of intelligent technology to solve the problem of traffic congestion becomes urgent needs.Firstly, this paper shows about the definition, classification and evolution law of traffic congestion, and classification, the basic parameters of traffic signal control and benefits evaluation index are expounded. At the same time, this paper points out the current world several mainstream signal control system and intelligent control methods.Secondly, the paper combines with the latest theoretical achievements in the field of artificial intelligence technology—multi-agent technology (Multi-Agent System, MAS), based on the control strategy of Multi Agent technology to research on city roads’ accidental coordinated control of traffic congestion. This paper mainly analyzed the characteristics of the Agent properties, structure model, at the same time, also probes into the distributed traffic control system structure based on Multiple Agent, combined with the advantage of the structure, points out that congestion coordination control feasibility and superiority of the introduction of Multi Agent technology.Again, combined with multi-Agent technology, this paper takes single intersection Agent for the control Unit, its composition and structure are analyzed. And focused on single junction Agent’s optimization of traffic signal control model which based on genetic algorithms (GA), the model is based on distributed traffic control system structure, the objective function defined as the intersection average delay of vehicles, using real-coded GA methods to achieve the single intersection congestion optimization control. Analysis of examples show that, after optimization of timing control method is superior to the present, control is better.Finally, the paper uses relevant theoretical knowledge of game theory, through analysis and description of problems between two intersections, explores a simple two junctions Agent’s game process and co-ordination mechanism. According to the actual traffic situation of the two intersections, comparing and analyzing performance index between the optimized and existed control method through an instance, which demonstrates feasibility and effectiveness of the coordination mechanism between Agents, also achieves a desired effects.