基于GA改进BP神经网络的建设工程投标报价研究  

Study of construction project bidding based on the BP neural network improved by GA

作　　者：王雪青[1] 喻刚[1] 孟海涛[1] 

机构地区：[1]天津大学,天津300072

出　　处：《土木工程学报》2007年第7期93-98,共6页China Civil Engineering Journal

基　　金：国家自然科学基金(60275020)

摘　　要：在建设工程项目投标报价过程中,如何来确定标高金直接关系到承包商能否中标和盈利以及施工企业今后的生存与发展。提出基于遗传算法(GA)优化BP神经网络的标高金预测方法。在分析BP神经网络基本原理的基础上,主要阐述了如何应用遗传算法来改进BP神经网络收敛速度慢和易于陷入极小值等缺点。对BP神经网络模型隐含层节点数进行优选后,建立起GA改进BP神经网络的标高金预测模型;最后应用该模型和一般BP神经网络模型对20个典型国际工程实例的标高金进行计算和预测。计算结果对比发现,经遗传算法改进后的BP神经网络模型在降低计算和预测平均误差的同时,迭代次数比一般BP神经网络模型也大大减少了。因此,该模型适用于求解如建设工程投标报价等非线性问题。During the bidding process of a construction project, how to calculate the mark-up directly influences whether the contractor can win the bidding and make profit, as well as the contractor＇s survival and prosperity in the future. A method for prediction of mark-up based on the BP neural network improved by GA （Genetic Algorithm） is proposed. On the basis of the basic theory of the BP neural network, discussions are provided on how to rectify the drawbacks of slow convergence and prone to convergence to minimum with the use of GA. After an optimal number of the hidden layers＇nodes of the BP neural network are selected, the prediction model of mark-up based on the BP neural network improved by GA is esatblished. The mark-ups of 25 typically international projects are calculated and predicted by using the proposed model and the normal BP network. The calculation results of the BP neural network improved by GA indicate that the average calculation and prediction errors are greatly reduced and the number of iterations is also smaller than that of the normal BP neural network.

关 键 词：BP神经网络 遗传算法 投标报价 标高金 

分 类 号：TU723.2[建筑科学—建筑技术科学]