A hillslope infiltration and runoff prediction model of neural networks optimized by genetic algorithm

Based on the measured data of hillslope simulated rainfall experiment in the Loess Plateau of China, the method of back-propagation neural networks optimized by genetic algorithms was used to establish the hillslope runoff and infiltration model. The rainfall intensity, rainfall duration, initial soil water content and slope were selected as the model inputs, the runoff volume and infiltration volume were the model outputs. Through of simulating and predicting, the results showed that simulation mean reletive errors were respectively 6.32% and 1.93%, the prediction mean reletive errors were 5.71% and 1.92%, respectively. In order to compare the prediction effects with other models, the unoptimized back-propagation neural network model and the Philip regression model under the condiction of fixed rainfall intensity were applied to predict the infiltration amount, the comprasion results showed the mean reletive errors of three models in infiltration amount prediction were separately 1.92%, 5.29% and 9.10%, the maximum mean reletive errors were separately 6.48%,25.88%, 20.36%, the prediction effects of optimized back-propagation networks had a better performance than the other two models obviously.