A Dragonfly Algorithm for Solving Traveling Salesman Problem

Traveling Salesman Problem (TSP) is considered as nondeterministic polynomial time hard problem. In the TSP, a salesman should visit a set of cities, and the distances between all pairs of cities are known in advance. The salesman has to find the shortest tour for visiting all cities exactly once and returns back to the starting city. Various methods have been used to tackle TSP, the most commonly employed methods are meta-heuristic algorithms. In this paper, TSP has been tackled by employing a newly created meta-heuristic algorithm, named Dragonfly Algorithm (DA), on well-known datasets (TSPLIB). The idea of DA has been inspired from swarm intelligence. To assess the quality of the proposed approach, it will be compared with other meta-heuristic methods that are available in the literature using the same datasets. The final results showed that the proposed DA-based TSP problem method is able to efficiently address the TSP and it produces competitively comparable results against others produces by well-regards comparative methods.