Hybrid Algorithm Based On Particle Swarm Optimization And Genetic Algorithm And Its Application In Function Optimization

Genetic algorithm is a kind of random search Algorithm, inspired by artificial life, andsimulated biological evolution process. The theory and application research of GeneticAlgorithm are widely attention by the majority of researchers, its application field also hasbeen widely promoted. solving function optimization problems, the choice of Controlparameters, crossover probability Pc and mutation probability Pm, had a great influence onthe performance of genetic algorithm. So far, generally accepted range of crossoverprobability Pc is between0.4and0.99, and mutation probability Pm is between0.0001and0.1,the rationality and the scientific of the two values range are lack of effective research. for thispurpose, in this paper, through a large number of experimental When get the global optimalsolution with the genetic algorithm, by the required number of iterations minimum as the goal,we mainly make a systematic study on the choice of the crossover probability Pc andmutation probability Pm for a class of functions, which can be expanded into a power series,by the analysis of the data of a large number of experimental, the conclusion is as follows:(1) Through the study and analysis of a large number of experimental data, When get theglobal optimal solution with the genetic algorithm, by the required number of iterationsminimum as the goal, we have obtained the optimal range of crossover probability Pc is0.6to0.99, mutation probability Pm is0.009to0.03;(2) Through large number of experimental data, we have made a systematic study forcross-impact between Pc and Pm, when choose the crossover probability Pc from the range ofthis paper suggested, the mutation probability Pm makes a remarkable difference for therequired number of iterations minimum When get the global optimal solution with the geneticalgorithm, on the contrary, when choose the mutation probability Pm from the range of thispaper suggested, the crossover probability Pc makes no significant difference for the requirednumber of iterations minimum When get the global optimal solution with the geneticalgorithm;(3) When get the global optimal solution with the genetic algorithm, to compare therequired computational between the mutation probability Pm value within the range proposedin this paper and generally accepted range, it has saved more than3times, to compare the required computational between the mutation probability Pm value within the range proposedin this paper and outside this interval value, it has saved more than4times.Particle Swarm Optimization (PSO) is an optimization algorithm, its principle is simpleand it is easy to operation realize. Algorithm was proposed by the wide attention of scholarsboth at home and abroad, various kinds of improved particle swarm algorithm had appeared.The convergence rate of particle swarm optimization algorithm is fast, and the solution ofparticle swarm optimization algorithm has memory function, but the global search ability ispoorer than genetic algorithm. This article proposes a new hybrid algorithm based on particleswarm optimization and genetic algorithm, combined with the advantages of two algorithmsrespectively, to foster strengths and circumvent weaknesses, Using hybrid algorithm based onparticle swarm optimization and genetic algorithm proposed in this paper, solve the functionoptimization problem of the standard test function, and perform an experiment with theparticle swarm algorithm and genetic algorithm, the experimental results verify that it is aneffective algorithm.