英文摘要 | Urban traffic congestion, traffic accidents and environment pollution have become increasingly prominent with the lagging of urban road construction resulting from rapid development of urbanization and sharp growth of cars. During the past few years, in order to alleviate these problems, Intelligent Transportation Systems (ITS) has been widely researched and applied, leading to some successful results. Nevertheless, the current ITS mainly faces two serious problems. On one hand, to conduct experimental analysis and evaluation on traffic control agents in real traffic system is considerably costly or even infeasible, while precisely establishing the model of city traffic system seems to be unlikely due to the strong nonlinearity, randomness and time-varying of city traffic system. On the other hand, as many kinds of traffic control agents exit, one control agent cannot meet the control demand of various traffic status, and therefore there is no way to put forward one optimal control agents combination for a regional traffic control system in order to obtain best performance. Parallel system theory and ACP method (Artificial systems+Computational experiment+Parallel systems), together with their application in solving traffic problems serve as effective solution to the above problems. Basically, ACP method is to establish the microscopic model for all the traffic entities with artificial systems, to evaluate and optimize the traffic control and management programs with computational experiments based on the executing results of the artificial systems, and at length to deploy the programs developed in the artificial system into actual system with parallel execution. The models, parameters, and experiments of the artificial system can then be adjusted according to the control and management performance of the actual system. Through iterative interaction between the artificial system and the actual system, reasonable traffic control and management programs can be ultimately achieved. However, it is difficult to ensure the instantaneity of the control system, being one bottleneck of this approach. There are two reasons. Firstly, when merely dependant on the traffic information at current time, the traffic control system cannot be intelligent, meaning that the control system cannot provide feedforward control. If the artificial system and actual control system send control command after collecting the current information from all intersections, the control in... |
修改评论