Optimization Design And Application Of Radial Basis Neural Networks Based On PSO Algorithm

Radial Basis Function Neural Network(RBF)has been widely used in various industries and fields because of its fast convergence speed,simple structure and ability to approximate any non-linear network.But in the practical application,the traditional algorithm can quickly build the network,but it is difficult to achieve the desired results,so it is often combined with the optimization algorithm to build the network,which is a hot spot in the current research of neural networks.In this paper,particle swarm optimization(PSO)is used to optimize the radial basis function(RBF)neural network and its model.As a parallel algorithm,particle swarm optimization(PSO)has attracted much attention in academia due to its advantages,and has shown its superiority in solving practical problems.At the same time,PSO algorithm also has some shortcomings,such as poor local search ability and early convergence.In solving complex problems,if the global optimization is not found,the particle may remain stagnant at a certain position;and in the later period,when the particle is near the extreme point,the search speed becomes slow,which will lead to the deterioration of the particle search ability.Scholars have proposed many improved algorithms,which have greatly improved the performance and efficiency of the algorithm.Then,it is still an important goal for researchers to develop algorithms with higher accuracy,efficiency and performance.This paper first introduces the radial basis function(RBF)neural network algorithm,deeply studies the principle,parameter setting and flow of RBF algorithm,then discusses the particle swarm optimization(PSO),studies its algorithm idea and the updated formula of standard algorithm,and finally chooses the particle swarm optimization(PSO)algorithm to optimize the RBF neural network.In the third chapter,after deeply analyzing the particle swarm optimization algorithm,aiming at the premature convergence and poor search ability of the particle swarm optimization algorithm,an improved optimization algorithm is proposed.By adjusting the inertia weight and learning factor,the early convergence problem of the standard particle swarm optimization algorithm is avoided,and the search accuracy and ability of the algorithm are improved.An optimized PSO-RBF neural network algorithm is proposed,which combines radial basis function(RBF)neural network with particle swarm optimization(PSO)algorithm.The improved particle swarm optimization(PSO)algorithm is introduced to optimize the center,standard deviation and weight of hidden layer basis function of RBF neural network,and a RBF neural network model based on PSO algorithm is established.In the fourth chapter,the optimized PSO-RBF neural network model is applied to water quality evaluation.Water quality evaluation is a complex process with characteristics of non-linearity,uncertainty and time-varying.In view of the low accuracy and robustness of the evaluation process,the PSO algorithm is used to optimize the RBF neural network to improve the accuracy and convergence speed of the neural network,and then to improve the accuracy of water quality evaluation.The validity of the proposed algorithm is verified by comparing it with the traditional one,and the superiority and reliability of the proposed algorithm are proved.