铁路路基沉降变形组合预测模型优化

由于传统灰色模型在预测波动性较大的数据时精度不高，提出一种改进的动态GM-Poisson-Markov组合预测模型。利用非等间距加权矩阵与无偏优化对灰色模型进行改进，通过原始序列的动态更新实现模型的参数更新，在此基础上与泊松曲线模型建立一种组合预测模型，并利用马尔科夫链进行残差修正，得到改进的动态GM-Poisson-Markov组合预测模型。利用汉巴南铁路路基沉降变形监测数据进行实例分析，将预测结果与泊松、灰色模型、非等间距无偏灰色模型以及组合模型预测结果进行对比分析，结果表明:模型对铁路软土路基沉降变形可取得较好预测效果，提高了预测精度与稳定性。

Optimization of Combined Prediction Model for Railway Subgrade Settlement Deformation

Since the traditional grey model is not accurate in predicting data with high volatility, this paper proposes an improved dynamic GM-Poisson-Markov combined prediction model. The gray model is improved by using non-equidistant weighting matrix and unbiased optimization, and the parameters of the model are updated through the dynamic update of the original sequence. On this basis, a combined prediction model is established with the Poisson curve model. The residuals are corrected to obtain an improved dynamic GM-Poisson-Markov combined prediction model by using Markov chain. Using the monitoring data of the subgrade settlement deformation of the Hanba South Railway to conduct an example analysis, the prediction results of the model in this paper are compared with the prediction results of the Poisson, grey model, non-equidistant and unbiased grey model and the combined model. The results show that the model can predict the settlement deformation of railway soft soil subgrade well and improve the prediction accuracy and stability.