Multi-swarm particle swarm optimization based on mixed search behavior

The paper develops a Multi-swarm particle swarm optimization (MPSO) to overcome the premature convergence problem. MPSO takes advantage of multiple sub-swarms with mixed search behavior to maintain the swarm diversity, and introduces cooperative mechanism to prompt the information exchange among sub-swarms. Moreover, MPSO adopts an adaptive reinitializing strategy guided by swarm diversity, which can contribute to the global convergence of the algorithm. Through the mixed local search behavior modes, the cooperative search and the reinitializing strategy guided by swarm diversity, MPSO can maintain appropriate diversity and keep the balance of local search and global search validly. The proposed MPSO was applied to some well-known benchmarks. The experimental results show MPSO is a robust global optimization technique for the complex multimodal functions.