Particle Swarm Optimization Algorithm Is Studied And Improved

Particle Swarm Optimization (PSO) is a new kind of evolutionary computation and was originally introduced by Kennedy and Eberhart in 1995. As a kind of swarm intelligence , it has proven to be a powerful global optimization method. PSO has been widely applied in function optimization and shows great potential in practice.However as PSO is a newly emerging optimization method, there are many research work should be substantiated. There are many improvements which increase the performance of the PSO, but they also increase the computation complexity of the PSO. So it is very important to seek an improved PSO which can greatly increase the convergence velocity of the classic algorithm and do not increase the computation complexity.Based on the analysis of convergence tendencies and confinements of the particle swarm, four methods are presented to improve the performance of the algorithm: increasing the convergence velocity, sustaining the diversity of the swarm, getting rid of the stagnation and using the advantage of other optimizer. Something important but easily neglected of the PSO are proposed, and their influences on the algorithm are also described.At last, the paper proposes a new Modified Particle Swarm Optimization (MPSO) , which dynamically classifies the Swarm Population according to whether the historical best position found by the particle is changed or not. Then according to the extent of the inheriting of the last iteration velocity, which the particle possesses, the MPSO is divided into three models (MPSO1, MPSO2, MPSO3). Five different benchmark functions were used to test the new MPSO, the results illustrate that the MPSO can greatly increases the convergence velocity of the algorithm and MPSO3 is more effective.