Research On Particle Swarm Algorithm Based On Good Point Crosser And Its Application

Particle Swarm Optimization is raised based on the combination of artificial life and swarm intelligence. It is evolution computing technology and is firstly proposed by Eberhart and Kennedy in the middle of the 90s of the century inspired by birds swarm, fish swarm and human society behavior rules.Because of PSO's merit, such as less parameter, easy algorithm implementation and quicker convergence rate, PSO is applied extensively in research and industrial production. But with the popularity of PSO's application, people find PSO's natural demerits, for example the slower convergence rate in the later period of algorithm, easily lose in local extremum and low precision and so on. To deal with these problems, many researcher start to research PSO with the propose of many improved algorithm. Among these, introduction some operators in genetic algorithm to PSO attract the most attention. The most famous is that Lovbjerg, Rasmuwsen and Krink in 2000 proposed that crossover operation in evolutionary algorithms (genetic algorithm) can be used in the PSO's HPSO model.Firstly the thesis introduce several improved PSO algorithm and standard PSO which are from basic PSO.Secondly, based on Crossover Algorithm and excellent characteristics of good point set, the theory proposed good point crossover PSO.In theory,the thesis shows the excellent algorithm performance of good point crossover PSO.Finally,the paper does experiment on five functions and two typical NP problem including TSP problem and Knapsack Problem to testify that good point crossover PSO is better than others in algorithm rate and precision.