非等时距DGM(1,1)模型及其在路基沉降预测中的应用

Subgrade Settlement Prediction of Expressway Based on the Improved Unequal Time-interval DGM(1,1) Model

传统路基沉降预测灰色模型往往采用将非等时距序列变换为等时距序列的方式来建立预测模型,建模过程较为繁琐。为此,本文另辟蹊径,通过引入新的序列累加累减生成运算,采用直接由非等时距序列构建预测模型的方式,以有效避免从非等时距序列到等时距序列转换的繁琐运算,提出了新的非等时距DGM(1,1)模型。在此基础上,应用最小二乘法获得新模型辨识参数值,并给出了其显著预测公式。最后,以某高速公路路基沉降预测为实例,说明了本文模型的有效性。

The traditional gray prediction model of subgrade settlement often was built by transforming the unequal time-interval sequence into time-interval sequence, whose modeling process was toilsome. Therefore, this paper put forward a new unequal timeNinterval DGM（1, 1） model through introducing a new sequence accumulation subtraction generating operation. This prediction model was built directly by the way of unequal time-interval sequence, which avoided complex calculations from the unequal time-interval sequence to the time-interval sequence transformation. On this basis, the identification parameters of the new unequal time-interval DGM（1,1） model by least square method was obtained, and its significant prediction formula was given. At last, a for example shows the effectiveness of certain highway embankment settlement prediction the proposed model.