Harmony search algorithm's parameter tuning for traveling salesman problem

This paper proposes statistical methods to find the parameter setting of an artificial intelligence technique, harmony search (HS) algorithm. The problem at hand is the travelling salesman problem (TSP) which is an NP-complete problem. Hence, a metaheuristic approach can give the near optimal solution in reasonable amount of computational time. The study makes use of the conventional HS to solve three benchmark problem sets in literature. The encoding and decoding schemes are presented. Then, the general full factorial design is used to find the HS' parameter setting. The analysis shows that HMCR and iteration number are significant. In addition, the appropriate setting of HMCR is 0.3 and iteration number is 5000.