摘要
重大基础设施工程是一类与国计民生息息相关的工程项目,因此,对于重大决策正确性的研究对于提升决策管理水平具有极其重要的现实意义。然而,由于重大工程往往以"孤本"形式存在,且涉及的人员、地域繁多,使得其研究常常具有历史数据资料缺乏、现场数据收集困难等特点。针对上述困难,结合港珠澳大桥项目的工程专家意见,基于贝叶斯网络原理,结合Expectations Maximization算法和Clique-tree Propagation算法,采用结构化系统开发方法,提出了一套用于重大工程设计方案的贝叶斯网络风险评估方法。案例部分,利用GeNIe 2.0软件平台为港珠澳大桥建设项目的主体工程部分构建了一个包含22个节点的三层贝叶斯网络。通过最后的敏感性分析、风险等级分析以及先后验概率比较,获得了很多有实际意义的结论,包括发现了SCP新技术、基槽开挖工期以及管节预制循环周期等7项关键风险。最后,探讨了关键风险因素的应对方法,并对结果中具有普世意义的风险进行了讨论与总结。
Mega infrastructure projects have a significant impact on the national economy and people＇s livelihood. The increasing scales and complexities of these projects pose a significant challenge to decision makers in risk management. However, the unique characteristics, stakeholder diversity, dispersed locations of a specific mega infrastructure project impose a difficulty in collecting historic data and on-site information. Based on the experience of engineer experts, we develop a structured Bayesian network model for project risk management based on maximum likelihood and Clique-tree propagation. Hong Kong- Zhuhai-Macao Bridge is used as an illustrative case to further show the use of the model. Multiple risk factors involving 22 nodes from 3 levels are obtained. Analysis results reveal seven key risk factors, such as SCP construction method, time delay of base slot excavation and the duration of prefabricated tubes. Finally, theoretic and practical implications are provided.
出处
《系统管理学报》
CSSCI
CSCD
北大核心
2018年第1期176-185,191,共11页
Journal of Systems & Management
基金
国家自然科学基金重大项目(71390521)
国家自然科学基金资助项目(91646123)
江苏省研究生科研创新计划资助项目(KYLX_0064)
关键词
风险管理
工程决策
贝叶斯网络
risk management
engineering decision making
Bayesian network
