A multi-population evolutionary algorithm based on constraint grouping for constrained multiobjective optimization problems

Constrained multiobjective optimization problems (CMOPs) are widely existed in the real-world applications and difficult to be solved due to the existence of multiple conflicting objectives and constraints. In the last few years, it has become a trend to design simple and effective helper problems to help solve the original CMOPs. However, most of the existing algorithms design auxiliary problems without considering the relationship between individual constraints and perform limited on solving different types of problems. To remedy this issue, a multi-population evolutionary algorithm based on constraint grouping, termed as CGMEA, is proposed. CGMEA creates independent populations to evolve single constraints in the first stage and analyzes the relationship between these populations. To fully utilize the relationship between single constraints, a constraints grouping method and an auxiliary population competition mechanism are proposed, respectively, to group the independent populations and allocate specific tasks for different groups. By designing more effective auxiliary populations evolutionary behavior, the diversity of both offspring generation and environmental selection are improved. Finally, the proposed algorithm and other six state-of-the-art algorithms are compared in detail on standard test suites and real-world applications. The results show the effectiveness of the proposed CGMEA.