基于CS、WLC、BPNN理论预测铁路工程造价的方法  

A Method of Estimating WLC of Scheduled Railway Projects Based on CS,WLC and BPNN Theorems

作　　者：段晓晨[1] 余建星[1] 张建龙[2] 

机构地区：[1]天津大学建筑工程学院,天津300072 [2]中国人民大学财政金融学院,北京100872

出　　处：《铁道学报》2006年第6期117-122,共6页Journal of the China Railway Society

基　　金：国家自然科学基金项目(70373032)

摘　　要：前期可行性研究投资估算和设计阶段概预算不准确是目前铁路工程“三超”的主要原因之一,在这两个阶段中,许多不确定因素影响着项目造价,并且这些影响因素与造价之间呈高度非线性关系,传统的投资估算和设计概预算方法难以对新建铁路项目造价进行准确计算。本文依据全生命周期造价WLC理论进行拟建项目造价测算和投资、设计方案优化,根据显著性成本CS理论提出显著性成本项目CSIs测算方法简化投资预测程序,并依据BP神经网络在大量已完工程资料中“提取”类似CSIs和显著性因子csf,从非线性角度实现了对项目投资的准确预测,并进行算例分析,从算例可以看到,预测值与实际值的相对误差很小,满足投资预测要求。The inaccuracy of cost estimation in the feasible research and design stage is one of the main reasons that causes investment of railway construction projects out of control in China. In those two stages cost estimations are affected by many uncertain factors, and the relationship among them is nonlinear and the traditional model is hard to solve them. This paper applies the CS（Cost-significant） theorem to simplify cost estimation, and sets up the forecasting methods of CSIs（Cost-significant Items）of WLC（Whole Life Cost）. Then the BPNN （Back-propagation Neural Network） model is made up according to the BP algorithm to ＂distill＂ CSIs and csf （cost-significant factor） from the data and information of completed projects, which provides a practical solution for those problems according to the nonlinear theory. The basic theories of BPNN, CS and WLC are introduced and their applications are illustrated with an example. From the example, we can see that the relative errors are so small that they can meet the accuracy demand of cost estimations after simulation, and the test result shows that the model based on the BPNN, CSIs and WLC theorems is accurate and successful.

关 键 词：投资估算 WLC CS CSIs BP神经网络 投资预测 

分 类 号：U212.32[交通运输工程—道路与铁道工程]