Improved RBF Neural Network for Nonlinear Identification System

Standard particle swarm optimization (SPSO) algorithm was modified by escape strategy of the particle velocity, and an escape PSO (EPSO) was proposed to overcome the shortcomings of being trapped in local optima because of premature convergence. To enhance the performance of radial basis function (RBF) neural network, the EPSO is combined with RBF neural network to form a EPSON hybrid algorithm. Compared with the hybrid algorithm of BP neural network (PSOBP), the experiment results show that EPSON has less adjustable parameters, faster convergence speed and higher precision in the nondifferentiable II. RADIAL BASIS FUNCTION NETWORK Artificial neural network (ANN) which is composed of a large number of neurons is a complex system of nonlinear dynamic and adaptive organization It has a highly nonlinear mapping capability RBF network is better than BP network in approximation capability classification and learning velocity. RBF was first proposed by Powell[9] and introduced into neural network literature. It has been widely used in pattern recognition, parameter identification, clustering analysis, etc. [10]. The topological structure of RBF network is commonly divided into three layers. The input node of the RBF network sends signals to the hidden layer and the output node is usually a simple linear function As input to output is nonlinear, and the network output is linear to adjustable parameter mapping the weight value of the network can be recursively calculated. The RBF network mathematica1 model is described as function identification.