Differential Evolution With Allocation of Mutation Strategy to Individual Based on Fitness Ranking
Jianyi Peng, Gang Chen, Xianju Li, Xuewu Han · Complexity · 2025
Real parameter single objective optimization has been the subject of extensive research. Differential evolution (DE) has exhibited remarkable performance. Recently, long‐term search has emerged as a new focal point of real parameter single objective optimization. In existing DE variants for long‐term search, integration of multiple mutation strategies or execution of local search is studied. In this paper, an algorithm named DE with allocation of mutation strategy to individual based on fitness ranking (AMSIFRDE) is proposed. In AMSIFRDE, the two aspects are both considered and enhanced. Different individuals are allocated to different mutation strategies, respectively, according to their ranking. In addition, a local search technique processes the median individual and the best one in turn in different generations. Experiments are conducted using the CEC 2020 and 2022 benchmark test suites and demonstrate that AMSIFRDE performs either better than or at least comparably to seven other algorithms for long‐term search.