Improved Particle Swarm Optimization Using Wolf Pack Search

Abstract

Particle Swarm Optimization (PSO) has excellent global exploration ability, but its local exploitation ability is not ideal. Wolf Pack Search (WPS), which is abstracted from the intelligent predatory behavior of the wolf pack, is an excellent local exploitation strategy and can be used to replace or improve the local exploitation capabilities of other heuristic algorithms. In order to improve the local exploitation ability of Particle Swarm Optimization without affecting the global exploration ability, a hybrid improved algorithm, named as WPS-PSO, based on wolf pack search is proposed. Though the simulation of fixed-dimension and multi-dimensional benchmark functions, and compared with the simulation results of the basic particle swarm algorithm, θ-PSO and Quantum Particle Swarm Optimization (QPSO), the results of 50 times simulations show that wolf pack search can improve the local exploit ability of PSO, and can better solve the multi-dimensional optimization problem.