Abstract:
To overcome the premature convergence of the standard particle swarm optimization (PSO) in solving multiple travelling salesman problems (MTSP), a new acceleration particle swarm optimization is constructed.Rely on the idea of mechanics, the movement of particle is described as search motion driving by force in solution space. The particle is attracted by personal best force, global best force and repelled by local best force.Thus the acceleration of particle depends on the resultant of forces. Using convergence criterions to estimate premature convergence, the local best will repel all the particles when premature convergence occurs, so the particle swarm can jump out the local best and continue to search. In order to improve the efficiency of the algorithm,a dimensional learning strategy of particle and a new coding method are designed for MTSP. The simulation results show that the proposed algorithm can effectively overcome the premature convergence, and improve the quality and stability of solutions.Thus it provides a feasible method for MTSP.
