OpenMP-Based Multi-core Parallel Cooperative PSO with ICS Using Machine Learning for Global Optimization Problem

Novel parallel cooperative multiple particles swarm optimization algorithm with immune clonal selection (ICS) using machine learning based on multi-core architecture is presented for global optimization problem in this paper, the proposed method named O-PCPSO-ICS. The O-PCPSO-ICS consists of one memory and bottom multiple swarms. In O-PCPSO-ICS, the global best individuals are saved into the antibody memory and promoted by using the improved ICS operator. An opposition-based learning operator is employed to accelerate the convergence speed of Pbests. Furthermore, excellent search information is spread among different subpopulations by a migration scheme. Finally, the proposed method is running on multi-core architecture using open multiprocessing (OpenMP). The numerical simulations validated the O-PCPSO-ICS has a better performance in global search, solution accuracy, and convergence speed. Meanwhile, the computational efficiency of the proposed method is greatly enhanced by parallelization.