Analysis and Improvement of Particle Swarm Optimization Algorithm

The effects of inertia weight on particle swarm optimization (PSO) performance are analyzed. A novel method of selecting inertia weight in PSO is developed, which can tune the expectations of inertia weights adaptively when the inertia weights are randomly selected and lead to effectively balance between the local and global search ability. Results of the two benchmark functions indicate that the PSO algorithm based on the strategy of random inertia weight (RIW) has been significantly improved on both optimization speed and computational accuracy.