Modified Self-adaptive Brain Storm Optimization Algorithm for Multimodal Optimization

Abstract

Multimodal Optimization is one of the most challenging tasks for optimization, since many real-world problems may have multiple acceptable solutions. Different from single objective optimization problem, multimodal optimization needs to both find multiple optima/peaks at the same time, and maintain these found optima until the end of a run. A novel swarm intelligent method, Modified Self-adaptive Brain Storm Optimization (MSBSO) algorithm is proposed to solve multimodal optimization problems in this paper. In order to find potential multiple optima, a modified disruption strategy is used for BSO algorithms to maintain the identified optima until the end of the search. Besides, the self-adaptive cluster number control is applied to improve Max-fitness Clustering Method with no need for a predefined subpopulation size M. Eight multimodal benchmark functions are used to validate the performance and effectiveness. Compared with the other swarm intelligent algorithms reported in the literature, the new algorithm can outperform others on most of the test functions.