Evaluation of Maintenance Policies Using a Two-Stage Pythagorean-Based Group Decision-Making Approach

The key to achieving the safety and reliability of the production system is to implement an efficient maintenance management system to reduce or eliminate equipment failures. Maintenance planning costs constitute a large part of operating costs. Determining the optimal maintenance policy plays a significant role in improving system reliability and availability as well as providing cost-savings. For this reason, it is necessary to determine and implement the most appropriate maintenance planning for the production system to ensure the reliable, smooth, and healthy operation of the system. The decision-making process is multidimensional and complicated due to the nature of the maintenace policy selection problem, so it involves various conflicting evaluation criteria in identifying the optimal maintenance policies. Multi-criteria decision-making (MCDM) instruments are appropriate for such complex issues. In this study, a hybrid-two-stage-MCDM-approach consisting of Pythagorean fuzzy Analytical Hierarchical Process (PF-AHP) and Pythagorean fuzzy TOPSIS (PF-TOPSIS) are carried out to determine the best maintenance policy, which is one of the most serious issues, both tactically and operationally, handled by one of the largest food companies in Turkey. Five main evaluation criteria (reliability, safety, cost, added-value, and feasibility), and six alternative possible maintenance policies (reliability-centered, predictive, time-based preventive, condition-based, opportunistic, and corrective maintenance) are considered to specify the optimal maintenance strategies. As a result of the application, the most appropriate maintenance strategy was emerged as reliability-centered maintenance (RCM). Other alternatives were appeared as predictive (PdM), time-based preventive (TBPM), condition-based (CBM), opportunistic (OM), and corrective maintenance (CM). The validation of the method was performed using the PF-VIKOR technique as a comparative study. In the final stage of the study, a sensitivity analysis based on criteria weights was conducted to test the robustness of the proposed integrated methodology.