A Multicriteria Decision-Making Model Based on the Integration of AHP and Bayesian Decision Trees for Selecting a Method for Replacing Submarine Pipes

Abstract

The X–Y Platform subsea pipeline owned by PT PHE OSES has been operating beyond its design life and has experienced significant degradation, as evidenced by 17 leakage incidents and the installation of nine repair clamps over the past five years. This condition requires strategic asset management decisions that consider not only costs but also technical and safety risks. This study aimed to develop a hybrid decision-making model integrating the Analytic Hierarchy Process (AHP) and the Bayesian Decision Tree (BDT) to determine the optimal pipeline replacement method. In this model, AHP was used to determine the priority weights of five main criteria: Cost, Health, Safety, Security, and Environment (HSSE), Technical, Flow Assurance, and Integrity. Subsequently, the BDT method was applied to combine these weights with historical failure probability data and posterior probability updates to calculate the utility values of four decision alternatives: Full Replacement, Segmented Replacement, Localized Repair, and Do Nothing. The results showed that the Integrity criterion received the highest priority weight (34%). Based on the posterior Expected Utility (EU) evaluation, Full Replacement was identified as the optimal alternative, with the highest EU value of 60.55, outperforming Segmented Replacement (52.23), Localized Repair (43.99), and Do Nothing (30.54). This integrated model provides objective technical recommendations to support management decision-making in effectively mitigating operational risks associated with subsea pipelines under conditions of uncertainty.