摘要
针对标准BP神经网络建筑工程项目投资估算模型收敛速度慢、预测精度低的问题,提出融合改进天牛须和正余弦双重优化算法(BAS-SCA)优化BP神经网络的建筑工程项目投资估算模型。以某市高校建筑工程项目为研究对象,分析相关文献并结合显著性理论初步选择工程造价影响因子,利用粗糙集属性约简算法筛选出关键因素;基于此,通过构建基于BAS-SCA-BP的神经网络估算模型实现快速、准确的建筑工程投资估算。研究结果表明:基于BAS-SCA-BP的估算模型较标准BP神经网络估算模型的估算精度有了大幅提高,与其他智能算法改进的BP神经网络估算模型的性能相比较,该模型在稳定性和预测精度方面表现更佳。
A construction project investment estimation model based on a optimization algorithm combining improved beetle antennas search and sine cosine(BAS-SCA)was proposed to optimize BP neural network. This model can overcome the slow convergence rate and low prediction accuracy of the standard BP neural network investment estimation model for construction projects. The construction project of a university in an anonymous city is taken as a research case. The relevant literature is analyzed,and the impact factors of project cost are preliminarily selected based on the significance theory. Then, the key factors are screened out by the rough set attribute reduction algorithm. By this, a neural network estimation model based on BAS-SCA-BP is constructed to achieve fast and accurate construction investment estimation. The results show that the estimation accuracy of the BAS-SCA-BP estimation model is greatly improved compared with the standard BP neural network estimation model. Compared with the performance of the BP neural network estimation improved by other intelligent algorithms,the proposed model is better in stability and prediction accuracy.
作者
雍秀珍
黄山
袁维
YONG Xiu-zhen;HUANG Shan;YUAN Wei(School of Civil Engineering,Lanzhou Jiaotong University,Lanzhou 730050,Chin)
出处
《工程管理学报》
2022年第5期136-141,共6页
Journal of Engineering Management
关键词
高校建筑工程
粗糙集理论
BAS-SCA
投资估算
university construction engineering
rough set theory
BAS-SCA
investment estimation
