Multi-population Co-evolutionary Multi-objective Optimization Algorithm Based on Improved Particle Swarm Optimization

Evolutionary computation is widely applied in multi-objective optimization due to its excellent global search capability and the characteristic of having multiple solutions. Among them, the Multi-Objective Particle Swarm Optimization (MOPSO) algorithm has received extensive attention from researchers because of its good development ability and fast convergence speed. However, a single population tends to fall into local optima when dealing with complex optimization problems. The algorithm still has issues such as low development precision and poor diversity of solutions. To address this, this paper proposes a Multi-Objective Optimization algorithm based on Multi-Objective Grey Wolf Optimization (MOGWO) and Improved Multi-Objective Particle Swarm Co-evolution (MCMOIPG). In each iteration, the population is randomly divided into MOPSO and MOGWO parts. In MOPSO, to balance the influence of solution precision and diversity, a learning factor combining non-dominated sorting and crowding distance is proposed. The position change of particles is influenced by the learning factor. The algorithm is compared with four other algorithms on the CEC2009 standard multi-objective test functions. Experimental results show that MCMO-IPG has superior performance in terms of solution precision, diversity, and algorithm stability.