Study on Nonlinear Prediction System of River Water Environmental Safety Based on Genetic Algorithm & BP Neural Network

This paper has presented a nonlinear model for the water pollution forecasting of river, which combines GA and BP neural network. Being not a simple model mixture, it can search the optimized BP network structure, learning efficiency eta, and additional momentum alpha based on its powerful searching capacity and then the ideal network about environmental pollution can be established using BP networkpsilas efficient ability to learn and constant training of network. It can effectively overcome those difficulties in determined the structure of BP nerve network, the learning efficiency eta and the additional momentum alpha.