Improvement Of Two Kinds Of Swarm Intelligent Algorithms And Its Convergence

Swarm intelligence algorithm is a heuristic algorithm based on collaborative social behavior of animals.Firefly algorithm and bat algorithm are both easy to operate and implement,so they have received extensive attentions from people.At present,these two algorithms have been used to solve practical problems,such as production scheduling,structural optimization.However,there are some problems two algorithms have: local optimality and low precision.So this paper is aims to improve two algorithms to enhance their performance and discuss the convergence of two algorithms.For firefly algorithm,an improved firefly algorithm based on gender difference(GDFA)is proposed.On the improvement of bat algorithm,an improved bat algorithm based on cellular automation(CABA)is proposed.The experimental results show they can improve the search ability and solution precision.In Chapter 1,the background and some improvement works of firefly algorithm and bat algorithm are given.The main contents of this paper is summarized.In Chapter 2,the elementary knowledge and update process of two algorithms are introduced,and the concrete implementation steps are given.In Chapter 3,the basic firefly algorithm is improved,and an improved algorithm based on gender difference(GDFA)is proposed.There exists gender difference between fireflies,which is different from hypothesis of basic algorithm.The population is divided into male and female subgroups.Then,two different update formulas are given according to their moving way.Male fireflies implement global search and female fireflies realize local search to balance the exploration and exploitation of algorithm.In order to further improve the search accuracy,the chaos search is conducted near the best position of the population.Finally,the convergence of GDFA is proved and the performance of algorithm is verified by comparison experiments.The results show that GDFA can improve the search efficiency and search precision.In Chapter 4,the basic bat algorithm is improved,and an improved bat algorithm based on cellular automation(CABA)is proposed.First of all,the idea of cellular automaton is introduced into the velocity update process,and a new neighborhood structure is introduced.Then,some good neighbors are used to replace the best individual to lead the fly direction,which can increase search scope of the algorithm.Next,a new local search equation is proposed to reduce the difficulty of parameter setting,and it is changed according to the number of iteration.Finally,the convergence of CABA and comparison experiments are given.The results indicate that CABA can effectively avoid premature convergence and enhance the search ability of the algorithm.