Research on Wavelet Neural Network modeling based on improved Particle Swarm Optimization algorithm

Abstract

For the shortcoming of Particle Swarm Optimization (PSO) algorithm in Wavelet Neural Network (WNN) training, a modeling approach of WNN based on improved PSO algorithm is proposed. The approach applied a PSO algorithm based on the strategies of multi-particle information sharing and self-adaptive inertia weight to optimize the parameters of WNN for modeling quality of WNN. The experiment result indicates that, compared with BP and Simple PSO (SPSO) algorithm in optimizing WNN, the approach had a better ability with features of convergence, precision, overcoming prematurity and local optimization, and was also a good method for nonlinear modeling.