A Novel Multi-Step Position-Selectable Updating Particle Swarm Optimization Algorithm
Yang Xiao-zong · Dianzi xuebao · 2009
Particle swarm optimization(PSO) algorithm is a new promising swarm intelligence optimization technology,and it has been extensively studied and applied because of its advantages of simpler theory,less parameters and better performance.However,each particle's individual minimum has a low updating rate,which has been one disadvantageous factor to affect this algorithm speed and precision.In this paper,we propose a novel multi-step position-selectable updating PSO algorithm.This algorithm decomposes the standard PSO velocity single-step updating formula into three steps and selects the best one among the three resultant positions as the final updated position.This scheme refines each particle searching trajectory,increases the updating speed of individual and global minimums,and consequently improves PSO algorithm converging speed and precision without increasing the computing complexity.Six classical testing functions,including Sphere,Rosenbrock and so on,are used to verify the proposed algorithm in two ways:a fixed iteration number test and a fixed time length test.Large numbers of simulations show that the proposed algorithm is simple,robust,and efficient,and meanwhile it outperforms other four existing classical algorithms.