The Research On Particle Swarm Optimization Algorithm Based On Multi-Swarm

As an important component of the swarm intelligence algorithm,the particle swarm optimization algorithm shines in the artificial intelligence field,and has become a hot topic of research by algorithm researchers around the world.Because of its advantages of fewer parameters,simple expression,easy implementation,and high efficiency,it has been applied to many practical optimization problems.However,in the face of complex multi-peak,high-latitude,multi-noise problems,the particle swarm optimization algorithm is easily trapped in the optimization process of local optimization,premature convergence caused by the problem of low convergence accuracy.In the thesis,a multi-swarm particle swarm optimization algorithm is the research object of the study,and the particle swarm optimization algorithm is apt to fall into the problem of local optimization.The main work is as follows:(1)Study the population size to multi-swarm particle swarm from the perspective of particle number and population number to optimize the performance of the algorithm,simulation experiments show that the more particles,the more populations,and the smaller the risk of the algorithm falling into a local optimum,the better the convergence results.(2)Study the effect of logical random grouping and physical location grouping on the performance of multi-swarm particle swarm optimization algorithm from the level of population division,and propose a multi-swarm particle swarm optimization algorithm based on clustering to achieve dynamic grouping.Simulation experiments show that the dynamic grouping of particle swarm optimization algorithm using clustering to achieve dynamic grouping process that is stable,it has feasible and effective.Compared with other excellent improved particle swarm optimization algorithms,it has better performance in global search ability,convergence speed and high solution accuracy.At the same time,the analysis of experimental results shows that the dynamic grouped multi-swarm particle swarm optimization algorithm needs to consume more computing resources during the optimization process.(3)The dynamic grouping of multi-swarm particle swarm optimization algorithm is applied to the cloud computing resource scheduling problem,analyzes the service provider's profit model,builds a fitness function,and the experimental results are analyzed to obtain a dynamic grouped multi-swarm particle swarm optimization algorithm.Outstanding performance in application examples.For the disadvantages of the particle optimization algorithm which can easily fall into local optimum,the work provides a reference for the particle population optimization algorithm to pre-set the number of particles and the number of populations.The proposed dynamic grouping multi-swarm particle swarm optimization algorithm has a good convergence precision performance that have a certain application value.