Exact Heuristic Algorithm for Traveling Salesman Problem

Abstract

Natural properties of stochastic searching strategies and operations in metaheuristic algorithms have important influence on convergence performance of various metaheuristic algorithms. Through similarity analysis to two kinds of metaheuristic algorithms, exact heuristic algorithm based on branch-and-cut is put forward according to change trend of similarity between two arbitrary high-quality solutions. In the meanwhile, two conditions were given in this paper because efficiency of branch-and-cut algorithm is closely allied to complexity of solved object. New heuristic algorithm can help metaheuristic algorithms finding superior solutions than other heuristic algorithms, and accelerate metaheuristic algorithms convergence. Simulation experiments show that new heuristic algorithm is efficacious.