Multi-objective Evolutionary Algorithm Based on Competitive Swarm Optimizer and Constraint Handling Techniques

A significant challenge in solving Constrained Multi-Objective Optimization Problems (CMOPs) is balancing convergence, diversity, and feasibility. Imbalance among these factors can prevent Constrained Multi-Objective Evolutionary Algorithms (CMOEA) from converging to the Constrained Pareto Front (CPF). When dealing with problems involving complex constraints and large objective spaces, most algorithms encounter difficulties. This paper proposes a novel Competitive Swarm Optimizer (CSO) with faster convergence and stronger search capabilities. To fully utilize infeasible solutions, a two-stage Constraint Handling Technique (CHT) is introduced, which leverages well-performing infeasible solutions to help the population escape local feasibility and explore feasible regions. To promote solution diversity, weak coevolution and probabilistic coevolution methods are employed during population evolution. Additionally, continual updating of the dual-archive further enhances solution convergence and diversity. Out of 23 test suites, Proposed algorithm obtained 13 of the best HV and IGD values, far more than any other algorithm. Simulation results on the LIRCMOP and DASCMOP test suites demonstrate the superiority of the proposed algorithm over other popular algorithms.