The Selection of Particle Swarm Optimization Learning Factors Values in Solving the Multiple Travelling Salesman Problem

Abstract

This paper addresses the question of whether fixed learning (acceleration) factors are an important factor in the Particle Swarm Optimization (PSO) by testing many selected values for those factors and apply them in solving the Multiple Traveling Salesman Problem (MTSP). Extensive experiments are done and those experiments show that the learning factors are problem dependent, therefore it is recommended to do the same experiments that are done in this paper for each problem that intend to be solved by PSO