基于误差反向传播算法的代建制项目风险评价研究

摘要

In the process of government investment project implementing agency construction, we must understand correctly the importance of risk management under agency construction and recognize the risk source,so as to assess the severity of project risk and the overall level of risk effectively which is an essential prerequisite of decision-making project. On the basis of expounding the importance of project under agency construction carrying on risk assessment,combined with the characteristic of project under agency construction and the generalization of domestic engineering practice,this paper set up common risk evaluation index system,put forward the BP neural network evaluation model based on error back propagation. Besides this paper uses the BP neural network evaluation model in the really construction to draw a conclusion that the model is feasible and hope the model can provide the decision support for risk prediction、risk response and risk aversion in the really construction,so as to improve the whole efficiency of the government investment project.%在政府投资项目推行代建制的过程中，必须正确认识代建制模式下的风险管理工作的重要性，认清风险的来源，有效评估工程项目风险严重程度和整体风险水平是决策项目上马的先决条件。本文在阐述代建制项目进行风险评价的重要性基础上，结合政府投资代建项目的特征和国内工程的实践进行归纳总结，建立了共性的风险评价指标体系，提出了基于误差反向传播算法的BP神经网络评价模型，并通过算例验证了该模型的可行性，为实际工程项目进行风险预测、应对、规避提供决策支持，从而提高了政府投资项目的整体效益。