优化的多变量变步长灰色模型及其在路基沉降预测中的应用

传统的单变量或等步长灰色预测模型，未考虑监测点间的系统关联性和实际沉降观测周期，往往具有非等间距的特点，不足以反映路基沉降变形的实际规律。在综合考虑实际工况中断面沉降点间相关性及不等时距观测的基础上，建立路基沉降预测的优化NMGM（1，n）模型。工程实例分析表明，优化的模型与建模工况相适应，具有良好的拟合及预测精度，能够满足工程实际应用需要。

Optimized Multi-variable Non-equidistance Grey Model and Its Application on Prediction of Subgrade Settlement

The traditional single variable or equidistant grey prediction model does not take into account the system correlation between monitoring points and the actual settlement observation cycle often has unequal interval characteristics, which is not enough to reflect the actual law of subgrade settlement deformation. In consideration of the correlation between the settlement point of each section of the actual condition and the non-equidistance observation, a optimized NMGM(1, n)model for prediction of subgrade settlement is established. The analysis of an engineering example shows that the optimized model is suitable for the modeling conditions. It has good fitting and prediction accuracy and can meet the needs of practical engineering applications.