Optimization and Application Analysis of Artificial Neural Network Based on Improved Particle Swarm Optimization

Abstract

To deal with the shortcomings of premature convergence and weak local search ability of Particle Swarm Optimization (PSO) algorithm, an improved PSO algorithm is proposed, based on a Lorenz chaotic sequence and weightself-adaptive adjustment. It puts forwards Hybrid Neural Network Model based on LSAPSO algorithm and RBF ANN (HANN), by optimizing the function center, expansion constant, network weights of RBF ANN with LSAPSO algorithm. It establishes a water quality evaluation model based on HANN, and conduct experiments on water quality evaluation with HANN forecast model. The results show that the model has good performance in water quality evaluation, forecast accuracy and correlation. And the HANN water quality evaluation and forecast is feasible and reliable, which can provide a new forecast method for many forecast fields.