A Dynamic Multipoint Detecting PSO

摘要

The chief aim of the present work is to propose a particle swarm optimization(PSO) by using a dynamic multipoint exploring approach. The main technique of this algorithm is that in the preceding phase of the algorithm, every particle can choose its searching direction and its moving velocity independently not being restricted or attracted by the optimal position of which have found by the parcle swarm and makes use of a dynamic multipoint random detecting method. It indicates, from the empirical results of four typical benchmark functions' optimization, that the optimization algorithm has the performance of rapid convergence rate, high accurate numerical solution, good stability and powerful robust. This proves that the algorithm is a promising means in solving the complex function optimization problems.