A mixed discrete particle swarm optimization for TSP

Particle swarm optimization (PSO) is a new evolution computing technology, its applications in discrete problem is a hot and valued field. Basing on the present algorithm of particle swarm optimization, this paper construct a special mixed discrete particle swarm optimization (MDPSO), and get a new formula of the velocity, via presenting the crossover and mutation of the genetic algorithm (GA). As applications, we propose it to traveling salesman problem (TSP) with 17 nodes and 14 nodes. Experimental results show that the proposed algorithm would greatly improve the efficiency of DPSO, especially, the convergence rate.