| The problems of shortest path and dynamic traffic assignment are important contents for identification, optimization and controlment research in transportation systems, especially for the analysis and optimized study of traffic network. In the large scale of traffic network, the demand that solve the real time problem of traffic system has not been satisfied yet while the current intellectualized algorithms ,which are well-developed swiftly, provide new methods to realize the shortest path and study dynamic traffic assignment issues.From what has been discussed above, Ant Colony Algorithm (ACA for short) has been introduced in this thesis. ACA is a new stimulant evolutionary algorithm put forward in1990s. This algorithm not only has predominant stochastic searching abilities, but also has self-adjustability and distributed trait, that is, it has characteristics of colony cooperation, positive feedback and parallelism. The processes of choosing route in ACA are similar with the process that vehicles select routes. Thereby, the ACA has been applied to solve the problems of shortest path and dynamic traffic assignment. To the disadvantages of easily getting into local optimum values and slowly solving problems, there have been elevated processes respectively, furthermore, the concrete realization steps and program of the algorithm are designed. Firstly, Adaptive Ant Colony Algorithm (AACA for short) is designed for the problem of shortest path. Then, to the problems of dynamic traffic assignment, the model of discrete dynamic users optimum assignment is found on the basis of the existing dynamic traffic assignment models combined with the characteristics of ACA. Thirdly, the Chaos Ant Colony Algorithm has been designed through the combination of ACA and Chaos Optimization Theory. At last, the rationality and validity is verified through simulation, and it provides a new method in solving the problems mentioned above with a certain academic reference and valuable practice.
|
PDF Full Text Request