Research On Improvement And Application Based On Grey Wolf Algorithm

The swarm intelligence optimization algorithm originated from people's simulation of some phenomena in the natural world,and simulated the biological evolution mechanism and physical laws in computer language.Compared with traditional optimization methods,the swarm intelligence optimization algorithm has strong adaptability in various complex optimization problems.These algorithms are now widely used in scientific research and engineering.Grey wolf optimization algorithm is a new swarm intelligence optimization algorithm in recent years.Because the grey wolf algorithm has the characteristics of simple parameter setting,it has been widely used in function optimization,intelligent control and other fields.In this paper,we propose an improved grey wolf algorithm.At the same time,this algorithm is used to solve the weight optimization problem of RBF neural network.The optimized RBF neural network has got good results in the classification of KDD CUP99 data set.Specifically,this study mainly includes the following two aspects:On the one hand,the grey wolf algorithm is improved for the disadvantages of the grey wolf algorithm.By introducing the index-based nonlinear convergence factor strategy and dynamic weight strategy to improve the grey wolf algorithm,the nonlinear convergence factor can dynamically balance the search ability of the algorithm,and the dynamic weights introduced can make the algorithm adapt to the environment dynamically during the convergence process and improve the convergence speed of the algorithm,using the benchmark test function to simulate the improved grey wolf algorithm,the experimental results show that the improved grey wolf algorithm is superior to the compared algorithm in convergence speed and convergence accuracy.On the other hand,the improved grey wolf algorithm is used in RBF neural network optimization.Firstly,a grey wolf algorithm based on the sine convergence factor is proposed for the disadvantage of low convergence accuracy of the grey wolf algorithm.Then the weights in the RBF neural network are mapped to individuals in the grey wolf algorithm to optimize the weights.The improved RBF neural network was applied to the classification experiment of KDD CUP99 data set.Experimental results show that the performance of the optimized RBF neural network on the data set is better than the comparison algorithm.