New Discrete Particle Swarm Algorithm for Traveling Salesman Problem

Abstract

In this paper, a novel algorithm called CIPSO for short based on particle optimization algorithm(PSO) and Chaos optimization Algorithm(COA) is presented to solve traveling salesman problem(TSP). We propose some new operators to solve the difficulties of implementing PSO into solving this discrete problem based on the special fitness landscape of TSP. Meanwhile embedded with chaos theory it can enhance particles’ global searching ability so as not to converge to the local optimal solution too quickly, and the introduction of information intercourse can enhance thire local searching ability. Comparing to SA, GA, ACS and so on, this new algorithm shows its validity and satisfactory effect on several benchmark test problems.