Research On Ant Colony Algorithm For Solving The Multi-Objective Traveling Salesman Problem

The multi-objective optimization traveling salesman problem(TSP)has many and complex criterion when evaluating the solution.No preference information,Pareto optimization in these solutions establishes a partial order,and the output of the algorithm results into a non dominated solution set rather than a single solution.In recent years,in order to solve the multi-objective traveling salesman problem which has proposed a variety of ant colony optimization algorithm.These multi-objective ant colony optimization algorithms(MOACO)according to the characteristics of multi-objective environment,have proposed many design ideas.The research on multi-objective ant colony algorithm mainly is concentrated on how to make use of optimization process to obtain the successful experience to further strengthen the positive feedback mechanism,but the method is difficult to avoid the premature convergence of the algorithm.In view of the problem,this thesis studies the failure lessons from the optimization process and applies them to the enhanced negative feedback mechanism.This provides a variety of information for the ants to effectively guide their foraging process.The work of this thesis is mainly reflected in the following three aspects:Firstly,the multi-objective ant colony algorithm is compared and analyzed by using box-line graph,H-index and Kruskal-Wallis test.Through the analysis,it is found that the performance of MOEA/D-ACO algorithm is better than other algorithms,and the different ways of pheromone update have a great influence on the algorithm.Based on the good algorithm(MOEA/D-ACO),the NMOACO/D algorithm is proposed by combining the negative feedback mechanism.The ant colony is divided into groups.A negative feedback pheromone matrix is applied at the same time.We apply for a positive feedback pheromone matrix in each group.The worst solutions of each iteration is used to update the negative feedback pheromone matrix of each group.By comparing the NMOACO/D algorithm with other known algorithms,the stability of NMOACO/D algorithm is not good.Aiming at this problem,an adaptive mechanism of SNMOACO/D algorithm is proposed.Because of the influence of different algorithms on the different factors,three kinds of pheromone updating strategies are...