改进GM(1,1)模型在路基沉降预测中的应用

GM(1,1)模型是沉降预测中应用较为广泛的模型之一，在实际应用当中发现其预测效果有时候较差甚至完全失效。为此从传统GM(1,1)预测模型建模原理出发，分析存在的理论缺陷，指出在形成预测公式时规定 ， 0.5为已知条件是不合理的，应当根据实际情况选用其他数据。通过引入平均斜率来计算 序列，并利用MATLAB程序进行数据处理，寻找最合适的 以及最佳维数区。某高速公路实测沉降数据计算表明，新方法有满意的拟合和预测效果，为提高建模精度提供了新的途径。

Application of improved GM (1, 1) model to settlement prediction of roadbed

GM (1, 1) model is one of the most widely used methods in settlement prediction, and it is found that the precision is low even defect. In this paper the structure principle of GM (1, 1) is analyzed and the defect is pointed out that and α should not be limited as the only know condition when forming the prediction formula. Instead, other data could be used. The average slope is used to calculate instead of constant 0.5. Program MATLAB is used to deal with the tested data and find the optimum and the optimum dimension. An actual example of an expressway is used, and it shows that the new model has satisfactory fitting effects and thus points a novel direction to heighten modeling precision.