A particle swarm optimization algorithm with empirical balance strategy

Abstract

As an important research direction of swarm intelligence algorithm, particle swarm optimization (PSO) has become a popular evolutionary method and received extensive attention in the past decades. Despite many PSO variants have been proposed, how to maintain a good balance between the exploration and exploitation abilities, and how to jump out of the local optimal position are still challenges. In this article, based on empirical balance strategy, a new particle swarm optimization (EBPSO) algorithm is presented. Firstly, based on an adaptive adjustment mechanism, the algorithm can choose a better strategy from two search equations, which can maintain the balance between the exploration and exploitation abilities. Secondly, to utilize the information of individual historical optimal solution and the optimal solution of the current population, a weight for adjusting their influence is introduced into the search equation. Thirdly, by introducing the diversity of population, a moving equation for dynamically adjusting the search ability of the algorithm is proposed. Finally, to avoid falling into local optimum and to search the potential location, a dynamic random search mechanism is proposed, which is designed by using the information of the current optimal solution. Compared with some state-of-the-art algorithms, the experimental results show that EBPSO has excellent solution quality and convergence characteristic on almost all test problems.