GM(1,1)-Logistic路基沉降组合预测模型研究与应用

Research and application of GM(1,1) and Logistic combination forecasting model for subgrade settlement

路基沉降预测是指导正确施工及运营期路基养护的一个重要因素。GM(1,1)模型及Logistic模型被广泛应用于路基最终沉降量的预测。基于组合预测的基本理论,结合GM(1,1)模型及Logistic模型的特点,提出了GM(1,1)-Logistic组合路基沉降预测模型,采用线性组合预测方法,以过去一段时间内组合预测误差平方和最小为原则来求2个预测模型的加权系数。结合工程实际监测数据的计算结果和分析表明,GM(1,1)-Logistic组合预测模型在预测精度上比单个模型具有更好的适用性。

Subgrade settlement forecasting is an important factor for the guidance of constructing correctly and subgrade maintenance in operation period. The GM （ 1, 1 ） and logistic model are widely used in subgrade settlement forecasting. Based on the basic theory of combination forecasting, according to the characteristics of GM （ 1, 1 ） and logistic model, a GM （ 1, 1 ） - logistic combination forecasting for subgrade settlement was proposed in this paper. Weighted coefficients of the two forecasting model computed by the principle which error square sum of combination forecasting was minimum in a past period. The results of engineering practical monitoring data show that the GM （ 1, 1 ） - logistic combination forecasting for subgrade settlement is more suitable than one forecasting model.