Abstract：Aiming at the problem of poor precision of two common settlement prediction of Logistic model and Gompertz model for predicting the final settlement of high fill body. Based on the commonness and individuality of Logistic model and Gompertz model, in order to give full play to their respective prediction advantages and complement their specific limitations, this paper optimizes the combination of the two models. Solving the optimal weighting coefficient with the objective function of minimizing the sum of squares of logarithmic errors of the combined model, and the optimal weighted geometric mean combination forecasting model is derived, so as to improve the confidence degree of the prediction accuracy of the settlement of the high fill.. The feasibility of the optimal combination forecasting model in engineering prediction is verified by fitting the monitoring data of specific high fill projects with each settlement prediction model. The results show that the precision and reliability of the optimal combination prediction model are better than any single model, and the adaptability is stronger. Especially, when the monitoring time exceeds the curve inflection point, it is almost close to the measured value. The combined prediction model can be applied to the optimization combination of multiple models, and can cover the advantages and disadvantages of each single model. As long as the weight ratio does not have negative value, the accuracy of settlement prediction can be greatly improved. so the combined forecasting model can be used as an effective method for predicting the settlement of high fill embankment in the later stage of engineering.
王博林 马文杰 王 旭 马学宁. 最优组合预测模型在高填方体沉降中的应用研究[J]. 土木工程学报, 2019, 52(S1): 36-43.
Wang Bolin Ma Wenjie Wang Xu Ma Xuening. Research on application of optimal combination prediction model in#br#
settlement of high fill body. 土木工程学报, 2019, 52(S1): 36-43.
