Research On Traffic Assignment Problems Based On Intelligent Optimization

With the development of world economy and people's living standards, vehicles as the people's main vehicle of transport are rapidly increasing, thus triggering the traffic congestion, traffic pollution and resource shortages have become increasingly serious. Advanced Intelligent Transportation Systems (ITS) as a main solution to traffic congestion has attracted more and more scholars, and traffic assignment as an important theoretical basis of Intelligent Transportation System has become the core of many scholars.Traffic assignment is to assign the traffic demands which are obtained by traffic forecast to the complex transport network with certain rules. In recent years, through the tireless efforts of many scholars, the study on the traffic assignment model is already quite mature, but the solution algorithm of traffic assignment can not meet the actual requirements of the transport network both in scale and efficiency. In this paper, for the problem that the computing time about traffic assignment problem is too unacceptable, the author improves the ant colony algorithm, genetic algorithms, particle swarm optimization algorithm and introduces them to the solve the problem of traffic assignment.First, introduces the basic theory about the traffic assignment problem, And for the phenomenon that the most popular impedance function(BPR function proposed by the U.S. Federal Bureau) can't satisfy the current situation of China, this paper presents a Segmented Impedance Function(SIF) which not only satisfy the actual situations but also easy to calculate. Secondly, based on the typical user equilibrium assignment model, the improved ant colony algorithm (ACO), genetic algorithm (GA), particle swarm optimization (PSO) are applied in a same traffic assignment problem, the results verify the three algorithm are availability. Furthermore, through analyzing the time and the stability of these algorithms, choose the most appropriate algorithm for solving the traffic assignment problem is the improved particle swarm optimization algorithm. Finally, the improved particle swarm optimization algorithm is applied in a dynamic traffic assignment network and the results are ideal. In addition, during the experiment, the Segmented Impedance Function(SIF) is also used in the traffic assignment network. Through comparing the results and BPR functions'results, we can conclude that the Segmented Impedance Function(SIF) can reduce the phenomenon of overflow greatly.