An Improved Circle Search Algorithm by Utilizing Multi-Strategy for Numerical Optimization

This is the authors' abstract. We don't add key points for this paper.

To address the disadvantages such as low optimization accuracy and falling easily into local optimality when Circle Search Algorithm (CSA) solving complex high-dimensional optimization problems, an improved CSA by utilizing multi-strategy (MSCSA) is proposed. Firstly, the Halton sequence is implied to population initialization to enhance the population diversity. Secondly, the adaptive differential mutation strategy is introduced to adaptively adjust the optimization range of the algorithm according to optimization process to improve the searching efficiency and optimization accuracy. Finally, the gradual opposition learning strategy is designed to help the algorithm to jump out of the local optimal. The results of CEC2017 test functions and Wilcoxon rank sum test show that MSCSA has better optimization accuracy and search efficiency than CSA and other compared algorithms such as Jellyfish Search Optimizer, Mutation-Driven Modified Grey Wolf Optimizer, and newly proposed Arithmetic Optimization Algorithm.