改进GM(1,1)模型在城市用水量预测中的应用

Application of Improved GM(1,1) in Forecasting City Water Demand

摘要: 本文在对GM(1,1)模型的不足之处进行了分析后，认为在对累加序列的拟合中，由k得到的离散解与累加序列之间存在较大误差，进而影响了模型的精度，所以提出了对k进行修正来提高预测精度，通过增加参数β，用梯度法求出β得到新的预测模型；最后将新模型应用到潍坊市用水量建模中，结果表明：本文提出的方法比之原有模型有较好的精度，由此拓宽了GM(1,1)模型的适用范围。

Abstract: After analyzing the deficiency of the traditional GM(1,1) forecasting model, the authors think that in the fitting accumulated sequence, the great error exists between accumulated sequence and discrete solution by k, and then has influenced the precision of the model. So this paper has pro-posed to amend k to improve the forecasting precision, through adding the parameter β and cal-culating the β by step method to get the new forecasting model. When applying it to the modeling of Weifang City water demand, the application result shows the method is effective and the application range is widened.