A Novel and Efficient Particle Swarm Algorithm for TSP

In this paper a novel Particle swarm algorithm is proposed for TSP. The key point of the proposed algorithm is the new and reasonable interpretation of the distance between two permutations that leads to new formulas for PSO. Indeed the relative order of cities in the permutations has a major role in defining the operators of the PSO. Many experiments were performed on Standard TSP library where all of the experiments showed the superiority of the proposed algorithm on the last PSO approaches for TSP. The proposed algorithm simply finds optimal solutions for TSP problems of sizes 70, where the last PSO approaches only reported their results on TSP with size of less than 20

کلیدواژه ها:

Discrete Particle Swarm Optimization ، Evolutionary Computation Algorithms ، Traveling Salesman Problem