BSO20: efficient brain storm optimization for real-parameter numerical optimization

Abstract

AbstractBrain storm optimization (BSO) is an emerging global optimization algorithm. The primary idea is to divide the population into different clusters, and offspring are generated within a cluster or between two clusters. However, the problems of inefficient clustering strategy and insufficient exploration exist in BSO. In this paper, a novel and efficient BSO is proposed, called BSO20 (proposed in 2020). BSO20 pays attention to both the clustering strategy and the mutation strategy. First, we propose a hybrid clustering strategy, which combines two clustering strategies, i.e., nearest-better clustering and random grouping strategy. The size of the subpopulation clustered by two strategies is dynamically adjusted as the population evolves. Second, a modified mutation strategy is used in BSO20 to share information within a cluster or among multiple clusters to enhance the ability of exploration. BSO20 is tested on the problems of the 2017 IEEE Congress on Evolutionary Computation competition on real parameter numerical optimization. BSO20 is compared with several variants of BSO and two variants of particle swarm optimization, and the experimental results show that BSO20 is competitive.