Urban Residential Water Demand Forecasting in Xi'an Based on RBF Model

Abstract

Based on the actual urban residential water demand of Xi'an, the Radial Basis Function (RBF) artificial neural network was used to forecast the urban residential water demand. RBF artificial neural network model was employed based on two input variables of population and Gross Domestic Product (GDP), one output variable of urban residential water demand. The performances in RBF forecasting of different spreads were compared and the forecasting result was the best when spread was 6. The urban residential water demand was forecasted for different influence factors, the variable of rainfall was eliminated. In order to get the performance of different models, some performance criteria such as Mean Error (ME), Root Mean Square Error (RMSE) and square of the correlation coefficient (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) were calculated for 2003-2005 testing data for RBF and Grey Model (GM). The urban residential water demands for different planning years were forecasted by RBF, GM(1,1) and the quota method respectively. The results indicated that RBF model was appropriate for forecasting the urban residential water demand.