The Research And Application Of The TSP Based On The Improved Artificial Fish Algorithm

Optimization problems need to be solved in many fields, and the fine solutions to the problems may lead to great economic benefit. With the increasing of the complexity and scale of the of the optimization problems, classical optimization methods which base on the strict modeling become difficult to be carried out. Artificial fish swarm algorithm is an animal’s autonomous method that bases on the principle of artificial intelligent.TSP problem is a typical problem in computer science, but also is a typical optimization problem of the composition. As a typical NP problem, TSP problem has become a standard that is used to test combinatorial optimization algorithms. Because of the broad application prospects of solving TSP, which also makes many researchers regard the TSP as an important research direction of the combinatorial optimization, and over the years TSP has become a hot topic of many scholars.The main purpose of this thesis is to improve the AFSA, and make the improved AFSA application to the TSP, which tests the superiority of the proposed algorithm comparing with the existing relatively good comparison algorithm.First, the background and research status of the basic AFSA was introduced, then the further research of the basic structure and optimization mechanism of the basic AFSA was done. And the various forming conditions of the artificial fish behavior and the artificial fish swimming mode in the algorithm was described. Detailed analyzes the structure and principle of AFSA. The analysis and explanation of convergence principle from some ways was given, such as fish swarm behavior and algorithm parameter. The simulation result was used to prove the excellent convergence of AFSA, detailed analyzes the major parameter of algorithm, for instance, visual and step. It discusses the effect of algorithm convergence performance from parameter.Secondly, the original AFSA were analyzed and summarized, and the way to improve of the algorithm was found out, then the methods and steps according to the AFSA’s defects were concluded. The improved algorithm called SSAFSA was given out based on the research of the other improved algorithm and validation. The adjustment of the targeted that includes global convergence, step length and vision as well as local minima problems was made out. The improved algorithm was compared with others through testing the verification function in many respects include the convergence speed, the precision optimization. The improved artificial fish algorithm effective and excellent was through the comparison of data.At last, the traveling salesman problem was solved by the improved AFSA, and joined the KNN algorithm in solving process to make the optimal path of solving TSP problems can be found faster and more accurate. The algorithm called SSAFSA-K can well solve the TSP problem was be proved based on classic database. From the comparison with others via the computing speed, the solution accuracy, the algorithm was proved as a feasible optimization algorithm for solving the practical problems.