Research And Application Of Swarm Intelligence Optimization Algorithm Based On Multiple Population Fusion

Swarm intelligence optimization algorithms are to simulate or reveal the certain natural phenomena and processes,which are developed from biological group of intelligent behavior.Since there is no requirement about continuous and differentiate for the optimization objective function,the result is not depend on the selection of initial values,and it has a global,parallel efficiency,strong robustness and generality,etc.Therefore,the study of swarm intelligence algorithm has important theoretical significance and practical value.This paper studied several kinds of typical swarm intelligence optimization algorithms in recent years for the global function optimization and applications.The specific works are described as follows:Firstly,based on krill herd optimization algorithm(KH),an improved krill herd optimization algorithm was proposed.By using the ideas of shuffled frog leaping algorithm memes grouping,it makes the krill herd algorithm more carefully,more comprehensively search.In order to verify the search performances of the improved algorithm,six benchmark functions are used to carry out the simulation experiments.Simulation results show that the parameters have great influence on the performance of KH-SFLA and the hybrid algorithm greatly increases the accuracy of solving the function optimization and optimization ability.Secondly,the flower pollination algorithm realizes the process of dynamic control conversion between global search and local search by the parameter p and the problem of balance between global search and local search are solved.Levy flight is adopted to make the algorithm have good global optimization ability.At the same time,five improved optimization algorithms(MFPA,MFPAS,CFPA,BFA-FPA and BFA-FPAS)are proposed.Multiple population algorithm has a strong ability of searching for the optimal solution.By simulation comparison,the performance is very stable.The inhibition rate can be effectively increased.At the same time,the accuracy of the solution is also increased significantly.Finally,on the basis of bat algorithm,its basic principle and algorithm flow are discussed.BA algorithm has the advantages of simplicity,fewer parameters,strong robustness and easy implementation,which has attracted great attention.Therefore,four improved algorithms of bat algorithm(IBA,SBA,LBA and CBA)are proposed and seven optimization algorithms(IBA,SBA,LBA,CBA,SBAS,LBAS and CBAS)are compared and analyzed.The simulation results show that the multiple population algorithm(IBA,SBAS,LBAS and CBAS)are better than the original algorithm in both the optimization ability and the convergence speed.The performance is relative-ly stable.All in all,the simulation results show that the parameters of the three-population intelligent algorithm are set and the function optimization results are good when combined with other algorithms.At the same time,the multiple population algorithm is superior to the original algorithm in performance.It is of great significance on solving the complex problems and their practical applications.