Large engineering project risk management using a Bayesian belief network Eunchang Lee a, * , Yongtae Park b , Jong Gye Shin c a Graduate Program in Technology and Management, Seoul National University, San 56-1, Shillim-Dong, Gwanak-Gu, Seoul 151-742, Republic of Korea b Department of Industrial Engineering, Seoul National University, San 56-1, Shillim-Dong, Gwanak-Gu, Seoul 151-742, Republic of Korea c Department of Naval Architecture and Ocean Engineering, Seoul National University, San 56-1, Shillim-Dong, Gwanak-Gu, Seoul 151-742, Republic of Korea article info Keywords: Risk management in large engineering projects Shipbuilding industry Bayesian belief network abstract This paper presents a scheme for large engineering project risk management using a Bayesian belief net- work and applies it to the Korean shipbuilding industry. Twenty-six different risks were deduced from expert interviews and a literature review. A survey analysis was conducted on 252 experts from 11 major Korean shipbuilding companies in April 2007. The overall major risks were design change, design man- power, and raw material supply as internal risks, and exchange rate as external risk in both large-scale and medium-sized shipbuilding companies. Differences of project performance risks between large-scale and medium-sized shipbuilding companies were identified. Exceeding time schedule and specification discontent were more important to large-scale shipbuilding companies, while exceeding budget and exceeding time schedule were more important to medium-sized shipbuilding companies. The change of project performance risks was measured by risk reduction activities of quality management, and strikes at headquarters and subcontractors, in both large-scale and medium-sized shipbuilding compa- nies. The research results should be valuable in enabling industrial participants to manage their large engineering project risks and in extending our understanding of Korean shipbuilding risks. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction Project risk management, one of the main subjects of project management (Raz & Michael, 2001), is the planning, organization, monitoring and control of all aspects of a project and it consists of risk identification, risk qualification, risk response development, and risk response control (Saynisch, 2005). Miller and Lessard (2001) pointed out that understanding and managing project risks in large engineering projects are challenging tasks at the early phase. The failure of large engineering projects has highlighted the importance of risk management mainly in the defense, con- struction and oil industries due to the serious damages that may be incurred (Williams, 1995). Active research has investigated pro- cess modeling and the methodologies of project risk management, in order to develop a systematic approach and integrated method- ology of project risk management (del Cano & de la Cruz, 2002; Raz & Michael, 2001). The use of diagrams such as cause and effect diagram and influ- ence diagram is one of the methodologies for project risk manage- ment. A diagram is suitable for the modeling of conditional probability relationships among risks, and is useful when handling complex problem. However, it is not easy to construct relation- ships and it is more complex than intuition-based analysis, so it has not been applied to project risk management as a widely used methodology (Han & Diekmann, 2001; Lyons & Skitmore, 2004; Raz & Michael, 2001; Simister, 1994). A Bayesian belief network is a graphical model that presents probabilistic relationships among a set of variables by determining the causal relationships among them (Heckerman, 1997). Because a Bayesian belief network constructs a cause and consequence dia- gram easily, it could be a suitable methodology for project risk management with systematic and integrated processes. Therefore, this study presents a project risk management procedure using a Bayesian belief network, applies this procedure to the Korean ship- building industry, and performs a project risk comparison between large-scale and medium-sized shipbuilding companies. 2. Literature review 2.1. Project risk management The main purpose of project risk management is to identify, evaluate, and control the risks for project success. The measure- ment of project success is difficult because it may be changed by project phase, and many stakeholders have different criteria to evaluate project success. However, the project success criteria are generally measured by time overrun, cost overrun, and technical performance (Baccarini & Archer, 2001; Williams, 1993). 0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2008.07.057 * Corresponding author. Tel.: +82 2 880 1380; fax: +82 2 872 8359. E-mail addresses: eunchang.lee@gmail.com (E. Lee), parkyt@cybernet.snu.ac.kr (Y. Park), jgshin@snu.ac.kr (J.G. Shin). Expert Systems with Applications 36 (2009) 5880–5887 Contents lists available at ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa