文献类型：期刊文献

英文题名：Chemical reaction-inspired dual-population co-evolutionary algorithm for many-objective optimization

作者：Ding, Weichao[1,2];Chen, Mingshan[1];Dong, Wenbo[1];Luo, Fei[1];Gu, Chunhua[1]

机构：[1]East China Univ Sci & Technol, Sch Informat Sci & Engn, 130 Meilong Rd, Shanghai 200237, Peoples R China;[2]Shanghai Key Lab Comp Software Evaluating & Testin, Shanghai, Peoples R China

年份：2025

卷号：289

外文期刊名：EXPERT SYSTEMS WITH APPLICATIONS

收录：;EI(收录号：20252218516199);WOS:【SCI-EXPANDED(收录号:WOS:001511695400006)】；

基金：Acknowledgement This work is sponsored by the Shanghai Pilot Program for Basic Research (22TQ1400100-16) , National Natural Science Foundation of China (No. 62403201) , and Natural Science Foundation of Shanghai (24ZR1415200, 23ZR1414900, 22ZR1416500) .

语种：英文

外文关键词：Dual-population paradigm; Chemical reaction algorithm; Many-objective optimization; Co-evolutionary computation

摘要：Many-objective optimization problems (MaOPs) refer to the multi-objective problems with more than three conflicting objectives that need to be optimized simultaneously, which are prevalent but more challenging in scientific research and engineering applications. Existing many-objective optimization algorithms mainly make improvements in environmental selection strategies but rarely focus on the generation of offspring, and balancing convergence and diversity is still very difficult. To address these issues, this paper employs four basic collision operators in chemical reaction optimization to generate offspring and proposes a novel chemical reaction-inspired dual-population co-evolutionary algorithm (called DPCRO), which aims to further balance convergence and diversity to improve the quality of solutions in solving MaOPs. The proposed DPCRO consists of two independent and parallel populations, which are the uni-molecular evolutionary population (UE) and the bi-molecular evolutionary population (BE). UE adopts the uni-molecular collision operators for individual convergence optimization and provides convergence information for BE, while BE uses the bi-molecular collision operators for individual diversity optimization and provides diversity information for UE. These two populations work collaboratively to effectively balance their own convergence and diversity through collaborative interaction mechanisms. We conduct performance comparison experiments using multiple benchmark test problems with different characteristics. The experimental results show that the proposed algorithm not only enhances the performance of solving low-dimensional multi-objective problems, but also has strong competitiveness in addressing MaOPs.