Risk management in shipbuilding using bayesian network with noisy-or

Abstract

Abstract This paper presents a method for risk management applicable in shipbuilding projects. We focused this study specifically to assess delay in project completion by analysing the production processes. The production processes were grouped into three major parts, namely the work of design, material procurement, and production processes. Risk data were collected from risk documentation of fast patrol boat constructions in PT. PAL Indonesia (the company). Bayesian network was used to assess the probability of project completion delay. Noisy-OR canonical model was used to define parameters needed for the Bayesian network model. Based on the current condition, the Bayesian network model predicted that there is a probability of delay of 0.67761739 that the project completion will be delayed. The Bayesian network model was also used to assess mitigation actions to reduce the probability of delay. Three mitigation action sets were provided based on the complexity of mitigation action: (1) less complex mitigation action, (2) more complex mitigation action, and (3) most complex mitigation action. The Bayesian network model predicting that the most complex mitigation action would provide the best result in mitigating the project completion delay.