基于BP神经网络的工程估价模型及其应用  

The Model and Application of the Cost Estimation of Construction Based on BP Neural Network

作　　者：叶青[1] 王全凤[1] 

机构地区：[1]华侨大学土木工程学院,福建泉州362021

出　　处：《厦门大学学报（自然科学版）》2008年第6期828-831,共4页Journal of Xiamen University：Natural Science

基　　金：福建省自然科学基金(2008J0196)资助

摘　　要：基于BP神经网络的工程估价模型具有高度的容错性和较强的泛化能力,通过对数据并行处理的方式能快速准确地估算出工程造价.本文根据BP神经网络原理,选取福建泉州地区的21组工程实例来建立模型,其中19组为训练样本,2组为检测样本,确定了13个主要造价影响因素作为网络的输入变量,工程造价作为网络的输出变量,经检验其精度符合工程投资估算和设计概算的要求.因此,用BP神经网络估算工程造价是行之有效的.The project cost estimation model, which was based on the neural network, was characterized by its high degree of fault -tolerance and strong generalization capability. This model can work out the project cost quickly and exactly through parallel processing of data. Based on the theory of BP neural network,this paper chose 21 examples of construction engineering in Quanzhou,FuJian to build up the model of project cost estimation. Among them,19 examples were used as training samples and 2 examples were used as test samples. According to the basic principles of the neural network and the characteristics of project cost estimation,thirteen major factors that affect the project cost were identified as neural network input variables, and project cost was identified as output variables. The test results showed that the precision meeted the requirements of project cost estimation and design estimates well, Therefore,the application of BP neural network is effective in project cost estimation.

关 键 词：工程估价 神经网络 估价模型 BP算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]