Application of the grey theory and the neural network in water demand forecast

Abstract

With the rapid development of society and economy of China, the imbalance between supply and demand is becoming increasingly conspicuous and toward further worsening. Therefore, the forecast of water demand is rather important for the reasonable planning and optimum distribution of water resources. There are various forecast methods for water demand. For a different city or region, a different and proper forecast method should be selected for the forecast. We take Yangquan City of Shanxi as the modeling object and respectively adopt grey forecast GM (1,1) model and RBF neural network model to forecast the water demand of Yangquan City in 1998, 1999 and 2000. For the grey forecast GM (1,1), the maximum relative error is −19.50%, the mean relative error is −13.85%.For the RBF neural network model, the maximum relative error is 2.29% and the mean relative error is 2.01%. The result indicates that the forecast precision of the RBF neural network model is better than that of the grey forecast GM (1,1) model and the forecast period is longer than that of the grey forecast GM (1,1) model. If both compared, the RBF neural network is more applicable for the forecast of the water demand of Yangquan City.