Parallel Particle Swarm Optimization Based on Parallel Model with Controller

Parallel computation is an effective method for solving complex large-scale engineering computation problems.A parallel PSO was studied based on the parallel model with controller that is coarse-grain parallel. It divides the whole group into several sub-groups. Every subgroup evolves in different processors independently and synchronously, exchanges and updates the best information periodically. The experiment results show that if the period of communication is selected appropriate, this parallel PSO not only has perfect speedup, but also improves the quality of result.