The Research And Improvement Of Particle Swarm Optimization

Particle Swarm Optimization is an adaptive stochastic optimization algorithm that based on the population search strategy. With the widely use of computer and the development of bionics, PSO is most important for solving optimization problems. The PSO algorithm has simple concept, high precision and fast convergence. Many scholars study on it from the PSO algorithm having been put forward. Considering the advantage of flexibility, stability, distributed control and self-leaning, PSO algorithm research has important science meaning and practical value either from the view of theoretical or applied research.Through the research on society behavior analysis of particle, an improved PSO algorithm was proposed based on a front particle's flying experience, and applied to solve the non constrained optimization problem. In the process of summary PSO algorithm improvement, the research finding that algorithm fusion has active significance, death penalty mixed with the PSO with constriction factor was proposed to solve the constrained optimization. The validity has been verified by the standard testing function.The first part introduces the research base of particle swarm optimization briefly. At the second part, instruction was given from the basic theory, parameters choose and topology structure. Improved PSO algorithms were listed by the purpose of improvement. The engineering applications of PSO algorithm was introduced.The third part gives the improved PSO and function test, it proves the availability of the algorithm.