A Mutation Triggering Method for Genetic Algorithm to Solve Traveling Salesman Problem

Abstract

This chapter reveals the effect of combining mutation operators in Genetic Algorithm. Though mutation is a very effective genetic operator to escape the local optima, it has an adverse effect on computation time. It is quite a challenge to say how much probability is appropriate for mutation. The major contribution of this chapter is to design a mutation triggering method by combining three mutation operators (Swap, Insertion and 2-Opt) and applying adaptive probability. To decide which mutation can be activated at a given generation, a decision-making method named Mutation Triggering Method is proposed. To measure the performance of the proposed method, the classical traveling salesman problem was considered. The experimental result shows that the proposed method was able to find a better solution (travel cost) than the other approaches. For computation time, the proposed strategy did as well as other mutation strategies. In conclusion, the combination of several mutation operators can ensure the benefits of diversity as well as the benefits of faster convergence.