摘要
GM(1,1)模型是城市用水量预测的一种有效的方法,但利用GM(1,1)模型难以反映序列的随机波动性。本文提出的平移变换和几何平均变换方法,不仅能构造更适合建立GM(1,1)模型的单调递增序列,也能有效地弱化原始序列的随机性,并保持其单调性,使其变化梯度趋于平缓。通过大连市2000~2006年用水量的预测结果表明,此方法能够反映出城市用水量所具有的波动特性,提高GM(1,1)模型的预测精度,可应用于对灰色振荡序列建立GM(1,1)模型,从而扩大了GM(1,1)模型的应用范围。
GM(1,1) is an effective method for the urban water consumption forecasting,but it is difficult to use GM(1,1) to simulate the random oscillation sequence.In this paper,the translation transformation and the geometric mean transformation are proposed,for they not only construct the monotonic increasing sequence which is more suitable for GM(1,1) model,but also weaken the randomness of the original sequence effectively.They maintain its monotony and make the change gently.The prediction results of Dalian water consumption in 20002006 show that this method can reflect the oscillation of urban water consumption effectively and improve the forecast accuracy for the oscillation sequence.Therefore,it expands the application scope of GM(1,1) model.
出处
《运筹与管理》
CSCD
北大核心
2010年第5期155-159,166,共6页
Operations Research and Management Science
基金
加拿大国际发展署(CIDA Tier 1)国际合作项目(S-61532)
碧流河流域GIS数字化管理中心示范工程(2005E21SF142)
关键词
灰色预测
GM(1
1)模型
振荡序列
用水量预测
平移变换
几何平均变换
grey forecast
GM(1
1) model
oscillation sequence
water consumption forecasting
translation transformation
geometric mean transformation