International Journal of Computer Theory and Engineering

Abstract

After the brief review of the basic principles and characteristics of Particle Swarm Optimization (PSO), a new particle swarm optimization, based on the simple evolutionary equations and the steep thermodynamical selection rule, are proposed to alleviate the premature convergence. The algorithm based on thermodynamical model, in which the selection rule simulates the competitive mechanism between energy and entropy in annealing to modify the exploitation and the exploration adaptively, can produce less off-particles in different free-energy scale not only to prevent the swarm from clustering and reduce the computational cost, but also to vary the diversity of the swarm and contribute to a global optimum output in the swarm. Relative experiments have been done; the results show the improved PSO performs very well on benchmark problems, and outperforms the other related PSO in search ability and stability.