摘要
Abstract
Forecast water demand is affected by many factors witch posses complex , high dimensional and nonlinear characteristics. Based on RBF and GRNN neural network algorithm principle,the paper constructed RBF and GRNN neural network water demand forecast model witch is used in the forecast of urban water demand , and combined the model with the basic BP neural network model and the gray GM (1,1) water demand model , compared and analyzed the fitting and forecast results. The results showed that ① The RBF and GRNN neural network model applied to water demand forecast is reasonable and feasible, the model possesses the strong generalization ability , high prediction precision and stable algorithm. Compared with the basic BP network algorithm, the model also possesses the advantages of fast convergence speed, less adjustable parameters and not easy to fall into local minimum value, and has a good application prospect; ② relatively speaking, the prediction accuracy of RBF and GRNN neural network model is better than that of basic BP network and gray GM (1,1) model.关键词
需水量预测/RBF神经网络/GRNN神经网络/BP神经网络/灰色GM(1,1)Key words
water demand forecasting/ RBF neural network/ GRNN nural network/ BP neural network/ greyGM(1,1)分类
建筑与水利
