沉降预测的非等步长灰色时变参数模型

Unequal step lengths grey time-varying parameters model for settlement prediction

根据灰色系统理论的信息处理原则,在灰色模型中引入时变参数,建立了沉降预测的非等步长灰色时变参数模型。模型充分考虑了预测系统的时变性和灰色性,在求解过程中,引入遗忘因子以修正预测结果,并根据后验方差比 C 确定遗忘因子 λ 的最优值,以提高预测精度。沉降预测实例表明:沉降预测的非等步长灰色时变参数模型具有较强的适用性,可降低沉降量的预测误差和提高沉降的预测精度。

The settlement prediction of soft ground is a typical time varying system, of which the parameters have time-varying property; hence, the time-varying property of parameters taken into account in the grey model for settlement prediction is significant. According to the information handling principle of the grey theory, the unequal step lengths grey time-varying parameters model for settlement prediction has been built. In the model, both time-varying and grey property are considered; and a forgetting factor, of which the optimum value is validated according to the posterior variance ratio c, is quoted in the model to correct prediction results. The example shows that the unequal step lengths grey time-varying parameters model for settlement prediction has applicability, and prediction error can be decreased and the prediction precision can be enhance by the grey time-varying paraheters model.