A Reinforced Self-Escape Discrete Particle Swarm Optimization for TSP

To deal with the problem of premature convergence and slow search speed of PSO, inspired by the classical 5-nearest neighbor method, a reinforced self-escape discrete particle swarm optimization algorithm (RSEDPSO) is proposed in this paper. The modified method of selecting candidate edges can enhance the performance of RSEDPSO to explore the global minimum thoroughly. The 5-relative nearest neighbor method introduced in this paper can produce candidate edges list more efficiently than the classical way, 5-nearest neighbor method. Experimental simulations indicate that RSEDPSO can not only significantly speed up the convergence, but also effectively solve the premature convergence problem.