基于PCA-PSO-BP神经网络的高速公路EPC项目成本风险研究

Research on Cost Risk of Expressway EPC Project Based on PCA-PSO-BP Neural Network

高速公路EPC项目建设周期长,不确定性、复杂性高,总成本控制难度大。为系统分析高速公路EPC项目面临的成本风险,提出基于粒子群优化算法和BP神经网络的EPC成本风险分析模型。首先,在识别成本风险因素的基础上,利用主成分分析法进行变量降维,简化网络输入节点;其次,通过粒子群算法优化BP神经网络,预测分析项目成本风险。结果表明,PCA-PSO-BP神经网络成本风险模型预测精确度高,可为高速公路总承包方评价项目成本风险提供一定的参考。

It is difficult to control the total cost that expressway EPC projects have a long construction period,high uncertainty and complexity.In order to systematically analyze the cost risks faced by expressway EPC projects,an EPC cost risk analysis model based on particle swarm optimization algorithm and BP neural network was proposed.Firstly,on the basis of identifying cost risk factors,principal component analysis was used to reduce the dimensionality of variables and simplify the network input nodes.Secondly,the BP neural network was optimized by particle swarm optimization to predict and analyze project cost risks.The results showed that the PCA-PSO-BP neural network model has high accuracy in cost risk prediction,which can provide a certain reference for the general contractor of expressway to evaluate the cost risk of the project.