Application of GA Optimized Wavelet Neural Networks for Carrying Capacity of Water Resources Prediction

The prediction of urban water demand using a small number of representative properties is fundamental in evaluating carrying capacity of water resources. Artificial neural networks (ANNs) have recently become popular tools in the prediction of urban water demand. In this paper, an iterative method which combining the strength of back-propagation (BP) in weight learning and genetic algorithmspsila capability of searching the satisfying solution is proposed for optimizing wavelet neural networks (WNNs). Taking the city of Hefei in China as an example, the proposed genetic algorithms optimized WNN that required a few representative properties as possible for input data is applied to predict urban water demand in the future several years. The prediction performance of the GA Optimized WNN is compared with traditional neural networks, and simulation results demonstrate the accuracy and the reliability of the prediction methodology based on the proposed model. Finally, urban water demand in Hefei, 2008-2010, is obtained which provide reference for coordinated development of socio-economic and water resources in Hefei.