龙学渊, 袁宗明. Frank-Wolfe算法在输气管道内腐蚀预测中的应用[J]. 油气储运, 2007, 26(1): 13-17. DOI: 10.6047/j.issn.1000-8241.2007.01.004
引用本文: 龙学渊, 袁宗明. Frank-Wolfe算法在输气管道内腐蚀预测中的应用[J]. 油气储运, 2007, 26(1): 13-17. DOI: 10.6047/j.issn.1000-8241.2007.01.004
LONG Xueyuan, YUAN Zongming. The Application of a New Algorithm of the Combination of Forecasts on Internal Corrosive Prediction of the Gas Pipeline[J]. Oil & Gas Storage and Transportation, 2007, 26(1): 13-17. DOI: 10.6047/j.issn.1000-8241.2007.01.004
Citation: LONG Xueyuan, YUAN Zongming. The Application of a New Algorithm of the Combination of Forecasts on Internal Corrosive Prediction of the Gas Pipeline[J]. Oil & Gas Storage and Transportation, 2007, 26(1): 13-17. DOI: 10.6047/j.issn.1000-8241.2007.01.004

Frank-Wolfe算法在输气管道内腐蚀预测中的应用

The Application of a New Algorithm of the Combination of Forecasts on Internal Corrosive Prediction of the Gas Pipeline

  • 摘要: 组合预测是对用多种预测方法进行预测的结果加权。建立了基于最小二乘法原理的组合预测模型,提出了求解此组合预测模型的一种新的算法,即Frank—w01fe算法,并将其应用于四川某输气管道内腐蚀速度预测的研究,应用结果表明,Frank-Wolfe方法较适用于求解组合预测问题的权重。

     

    Abstract: The combination forecasting is to weight results of the forecasting used by different methods. In this paper, a new algorithm of forecasts, the Frank-Wolfe algorithm, corresponding to a combinative forecast model based on the principle of OLS is presented. Its application to the forecasting of internal corrosive rate of gas pipeline shows that it has the highest precision in forecasting. Therefore, it is a simple and effective method of combinative forecasts in solving combinative forecast weight.

     

/

返回文章
返回