文献类型：期刊文献

英文题名：A large-scale multiobjective evolutionary algorithm with overlapping decomposition and adaptive reference point selection

作者：Gao, Mengqi[1,2];Feng, Xiang[1,2];Yu, Huiqun[1,2];Li, Xiuquan[3]

机构：[1]East China Univ Sci & Technol, Dept Comp Sci & Engn, Shanghai 200237, Peoples R China;[2]Shanghai Engn Res Ctr Smart Energy, Shanghai, Peoples R China;[3]Chinese Acad Sci & Technol Dev, Beijing 100038, Peoples R China

年份：2023

卷号：53

期号：19

起止页码：21576

外文期刊名：APPLIED INTELLIGENCE

收录：;EI(收录号：20232314199841);WOS:【SCI-EXPANDED(收录号:WOS:001002593700001)】；

基金：This work is supported by the National Natural Science Foundation of China (No.62276097), Key Program of National Natural Science Foundation of China (No.62136003), National Key Research and Development Program of China (No. 2020YFB1711700), Special Fund for Information Development of Shanghai Economic and Information Commission (No.XX-XXFZ-02-20-2463) and Scientific Research Program of Shanghai Science and Technology Commission (No.21002411000).

语种：英文

外文关键词：Large-scale multiobjective optimization; Overlap decomposition; Adaptive reference point selection; Ensemble optimization

摘要：Many large-scale multiobjective optimization problems with large decision space hinder the convergence search of evolutionary algorithms in various practical applications. Using the divide-and-conquer strategy to decompose the large-scale multiobjective problem into some subproblems and collaborative optimization is an effective strategy. However, the interactions between decision variables may cause many indirect interactions, which make complex high-dimensional problems impossible to decompose successfully using existing decomposition techniques. This paper proposes a multiobjective evolutionary algorithm with overlapping decomposition and adaptive reference point selection (MOEA-ODAR) for solving large-scale multiobjective problems. First, a decision variable overlap decomposition approach is suggested to group decision variables into several exclusive subcomponents. An adaptive resource allocation ensemble optimization method is then proposed to allocate corresponding resources to subcomponents with different structures. Finally, an adaptive reference point selection method based on Pareto shape estimation is designed to optimize the specific subcomponents. The theoretical analysis of the correctness of overlapping decomposition decision variables and collaborative optimization is presented. It is compared with the newly proposed excellent large-scale multiobjective optimization algorithm on many test problems. The experimental results show that the proposed algorithm performs better in terms of convergence, population distributivity, and computational efficiency. In addition, the superiority of the algorithm on large-scale many-objective problems is verified.