A Memetic-Clustering-Based Evolution Strategy for Traveling Salesman Problems

摘要

A new evolution strategy based on clustering and local search scheme is proposed for some kind of large-scale travelling salesman problems in this paper. First, the problem is divided into several subproblems with smaller sizes by clustering, then the optimal or the approximate optimal tour for each subproblem is searched by a local search technique. Moreover, these tours obtained for the subproblems are properly connected to form a feasible tour based on a specifically-designed connection scheme. Furthermore, a new mutation operator is designed and used to improve each connected feasible tour further. The global convergence of the proposed algorithm is proved. At last, the simulations are made for several problems and the results indicate the proposed algorithm is effective.