Two Different Methods For Forecasting Urban Water Demand

Urban water demand prediction is not only the basis of water resources planning and management,but also an important part of water supply systems.Apart from that,it is still an effective measure to release the conflict between the demand and supply of water resources,which can contribute to the sustainable utilize of water resources.However,due to the uncertainty of water demand data including imprecision in measurement and the uncertainty of influential factors such as natural environment,socio-economic development and related water policy,the prediction accuracy of the traditional models has been limited.Therefore,how to exactly forecast the urban water demand is worthy to be investigated.Taking into account the uncertainties within the framework of water demand forecasting,this dissertation attempt to combine the uncertain theory with traditional statistics methods to propose two different water demand prediction methods.In addition,we use the Xian and Beijing as two examples to illustrate our proposed methods are reasonable and effective.The main contribution of this dissertation are shown as follows.In the first part,the uncertain linear regression model with asymmetric triangular uncertain coefficients is built for the water demand prediction.Further,two parametric estimation methods are presented based on linear programming technique and nonlinear programming technique respectively.Then the domestic water of Tianjin city datasets as an application is adopted to compare the forecasting performance of the proposed two methods on the basis of classical measure methods.To further demonstrate the ability of the developed uncertain linear regression model to deal with the inaccuracy of information,the traditional regression method is selected as a competitor.The result shows that the proposed uncertain regression models outperform over the traditional regression model,especially for the uncertain regression model based on the nonlinear programming method.In the second part,the uncertain time series prediction method is introduced for the water demand prediction.Uncertain time series is a sequence of imprecisely observed values that are characterized by uncertain variables and the corresponding uncertain autoregressive model is employed to describe it for predicting the future values.First,by defining the auto-similarity of uncertain time series,the identification algorithm of uncertain autoregressive model order is proposed.Second,a new parameter estimation method based on the uncertain programming is developed for the case of imprecise observations,within the proposed method,the original problem including uncertain measure is transformed to the equivalent crisp mathematical programming under the certain confidence level.Third,the imprecise observations are assumed as the linear uncertain variables and a new ratio-based method is presented for constructing the uncertain time series.Finally,the proposed methodologies are applied to model and forecast the Beijing’s water demand under different confidence levels and compared with the traditional time series method.The experimental results show that the performance of the proposed method is better than the traditional time series method.The research will enrich the urban water demand prediction methodology and broaden the application of uncertain statistics.The result will provide support to the decision-making for urban water resources planning and optimal allocation.