采用灰色-马尔可夫耦合模型预测古树屋边坡变形

基于预测拟合值和实际监测值的相对误差划分不同的预测状态，利用马尔可夫链对灰色GM(1，1)预测值进行修正，计算得到边坡变形的范围和相应的预测值。以古树屋开挖边坡为研究对象，利用耦合模型对其变形进行预测，得到边坡的变形预测值为57.006 0 mm，预测的变形范围大小在56.845 8～61.755 4 mm之间。实例分析结果表明:与线性回归模型以及单纯灰色模型预测结果比较，采用灰色-马尔可夫耦合模型对开挖边坡变形进行预测的结果精度较好，且能较准确地预测边坡变形大小的范围。

Prediction of Gushuwu Slope Deformation with Grey-Markov Coupled Model

The predicted status is distinguished depending on the relative error between predicted fitting value and actual monitoring value,and grey GM(1,1)predictive value is corrected with Markov Chain and then the range of slope deformation and the relevant predictive value could be calculated.Thus the predicted value of the slope deformation is calculated to 57.006 0 mm varying from 56.845 8 to 61.755 4 in such manner that take Gushuwu cutting slope as the study object and predict the deformation with the coupled model.The case analysis shows that the accuracy of the prediction of cutting slope deformation with grey-Markov coupled model is higher than that with linear regression model and simple grey model;furthermore,the deformation range could be predicted accurately.