Improvement And Application Research Of Artificial Fish Swarm Intelligent Optimization Algorithm

With the expansion of industrial production,the optimization problem in industrial production also increases.After 1980 years,scientists have put forward many intelligent algorithms to solve this kind of problem.They take the life habit of nature as the research object,abstract a series of new ideas and methods to solve this kind of industrial production problems.Artificial fish swarm algorithm(AFSA)is put forward by the fish foraging behavior,group behavior and tailgating poly a swarm intelligence optimization algorithm.The algorithm has good global search ability,fast convergence speed,wide application range and so on.But AFSA also has many problems in solving the practical problems,such as slow convergence and low search accuracy.So this paper focuses on the improvement of the algorithm and two aspects in the application.The main research work is as follows:(1)Aiming at the problem that the AFSA can be easily trapped into random walks in a relatively flat area,and at the end of the algorithm,the convergence rate of algorithm reduce and the searching ability slow down,a new intelligent optimization algorithm with mutation operator and dynamic field is proposed,which has Inheritance the advantages of simple and easy to implement of AFSA,and it also can overcome the limitation of the high density of the blind random walk and the non-global-extreme point with high density fish,thus it can improve the performance of the algorithm in the operation efficiency and accuracy.(2)In view of the shortcomings of slow convergence speed and easy to fall into local extreme value of BP neural network,the improved,AFSA is applied to the learning process of BP neural network.By introducing mutation operator,which improved visual field and search step size,the AFSA algorithm has better global search ability and faster search efficiency.By using of the improved AFSA algorithm in the training of the weights and thresholds in BP neural network,the training speed and precision are improved.In a word,based on a comprehensive analysis of AFSA,the paper puts forward the corresponding optimization method for some problems of AFSA.The improved AFSA is applied to optimize the BP neural network.In the end,the paper summarizes the work in the paper.