The analysis of strategy for the boundary restriction in particle swarm optimization algorithm

Abstract

Particle swarm optimization has been applied to solve many optimization problems because of its simplicity and fast convergence performance. In order to avoid precocious convergence and further improve the ability of exploration and exploitation, many researchers modify the parameters and the topological structure of the algorithm. However, the boundary restriction strategy to prevent the particles from flying beyond the search space is rarely discussed. In this paper, we investigate the problems of the strategy that putting the particles beyond the search space on the boundary. The strategy may cause PSO to get stuck in the local optimal solutions and even the results cannot reflect the real performance of PSO. In addition, we also compare the strategy with the random updating strategy. The experiment results prove that the strategy that putting the particles beyond the search space on the boundary is unreasonable, and the random updating strategy is more effective.