基于改进灰色GM(1,1)模型的高速公路路基沉降预测

Prediction of expressway subgrade settlement based on animproved gray prediction model

有效预测高速公路路基沉降的发展趋势,可以提高道路维修效率,降低事故发生率。根据路基沉降影响因素和特性,提出一种改进的GM(1,1)预测测算法。采用自动寻优定权的方法优化传统GM(1,1)模型的背景值,采用最小二乘法优化初值,有效克服背景值与初值选择的盲目性,并能保证较高的精度。分别对长湘高速某段和垄茶高速某段路基沉降进行预测试验,本文算法预测误差均值分别为2.12%和1.38%,方差比为0.137和0.146,精度等级达到一级。

The effective prediction of the development trend of expressway subgrade settlement can improve road maintenance efficiency and reduce the accident ratio. According to the subgrade settlement factors and features,this paper presents an improved forecasting method. The automatic optimization method was used to chose the background value and the least square theory is used to improve the initial value. In the prediction test of two examples,the results indicate that the means of the relative errors are 2. 12% and 1. 38%,the ratios of mean-square are 0. 137 and 0. 146,and the precision grade is I.