论文部分内容阅读
路基沉降预测一直是道路工程领域的研究重点和难点。常用的GM(1,1)模型所预测的路基沉降值精度相对较低,特别当数据序列急剧变化时,GM(1,1)模型的误差值可能会更大甚至失效。针对传统GM(1,1)模型存在的问题,通过改变初始值,增加扰动因素β优化初始条件。同时利用非齐次指数函数拟合模型中变量的一次累加生成序列优化背景值,提出了初始条件和背景值双优化的新GM(1,1)模型。通过MATLAB软件编程实例计算表明,双优化之后的新GM(1,1)模型较原模型相比,其预测精度有了较大幅度的提高。
Subgrade settlement prediction has been the focus and difficulty in the field of road engineering. The accuracy of the subgrade settlement value predicted by the commonly used GM (1,1) model is relatively low, especially when the data sequence changes drastically, the error value of GM (1,1) model may be larger or even invalid. Aiming at the problems existing in the traditional GM (1,1) model, the initial condition is modified by changing the initial value and increasing the disturbance factor β. At the same time, a new GM (1,1) model with double optimization of initial conditions and background values is proposed by using one-time accumulation of variables in the nonhomogeneous exponential function fitting model to generate sequence optimization background values. The calculation example of MATLAB software shows that the new GM (1,1) model after double optimization has a much improved prediction accuracy compared with the original model.