Multi-core based parallelized cooperative PSO with immunity for large scale optimization problem

Abstract

A parallelized cooperative multiple particles swarm optimization algorithm with immunity mechanism based on the multi-core architecture is proposed for large scale optimization problem in this paper, named M-PCPSO-I. A novel memory information sharing scheme is designed for particles and facilitates communication among different swarms in the population space. The global best individuals selected from sub-swarms are saved in the leader set and promoted by using the improved immune clonal selection operator. The M-PCPSO-I algorithm is paralleling implementation on a share-memory computer system through the multi-core architecture. The high dimension problem results validated the proposed algorithm have good computational performance, and also the computational efficiency is greatly enhanced by multi-core parallelization.