An improved harmony search for travelling salesman problem
Shih-Pang Tseng · 2016
Harmony search(HS) is a music-inspired metaheuristic for optimization problem. HS is simple to implement and apply on several kinds of problems. But HS is not successful to solve large size of travelling salesman problem(TSP). In this paper, we try to find the difficulties of HS on TSP. And we propose the elitists harmony search(EHS) to overcome these difficulties. The EHS reorganizes the harmony memory, which is divided into two groups: elitists and explorers. Elitists are used to raise the intensification and explorers are for the diversification. The experimental result shows that EHS is better than HS average 25.49%, and EIHS is better than IHS average 20.09%. The difference of execution time is very small. And for larger TSP benchmark, the EHS is more better than HS.