基于非平稳时序的城市用水量ANN-ARMA预测模型.doc

基于非平稳时序的城市用水量ANN-ARMA模型 (1北京理工大学 北京 100081) : 目的：面对城市水资源供需矛盾日益加剧的现状，城市用水量预测已成为城市建设与水资源规划工作的重要内容之一。它直接关系到城市水资源的可持续利用和社会经济的可持续发展。目前，基于平稳时间序列的城市用水量短期预测方法研究比较广泛，并已经取得了较好的预测效果；但是，中长期的城市用水量由于受到社会诸多因素的综合影响，其时间序列具有明显的趋势性和随机性，故预测方法复杂，且研究相对较少。该论文研究了基于非平稳时间序列的城市用水量中长期预测方法。 方法：根据时间序列分析的有关理论与方法，即非平稳时间序列分解为确定项和随机项两个部分提出了集成人工神经网络（ANN）与自回归滑动平均（ARMA）的模型。非平稳时间序列该模型人工神经网络模型拟合确定项非平稳变化趋势；自回归滑动平均拟合随机项以表示平稳随机成分。两模型的预测值之和作为的值。与实际的相对误差不超过%。模型预测人工神经网络模型预测人工神经网络与自回归滑动平均预测；非平稳时间序列；人工神经网络；自回归滑动平均模型ANN-ARMA model for forecasting urban water consumption based on non-stationary time series CAI Feng，ZENG Feng-zhang (School of Management and Economics, Beijing Institute of Technology, Beijing, 100081) Abstract: Purposes: In the current situation that contradiction between supplies and demands of urban water resources is more and more intense, forecast for urban water consumption has become one of important works regarding urban construction and water resources planning. It directly affects sustainable utilization of urban water resources and sustainable development of social economy. At present, it is very wide that studies on methods of short term forecast for urban water consumption based on stationary time series, and they have gained better effects. Because time series of urban middle-long term water consumption that is affected by lots of social factors, are characterized with tendency and randomness, forecast methods are more complicated and fewer are researched. This paper researched the method of middle-long term forecast for urban water consumption based on non-stationary time series. Methods: In terms of concerned theories of time series analysis, i.e. non-stationary time series can be divided into the certain part and stochastic part, a forecast model of integrating Artificial Neural Network (ANN) with Auto Regressive Moving Average (ARMA) is presented. Aiming at non-stationary time series of urban middle-long term water consumption, their certain part denoting non-stationary trend can b