Managerial decision making for complex service systems Optimization

The present paper deals with managerial decisions for Predictive Maintenance (PrdM) of complex service systems. We propose a Multi-Criteria Decision-Making (MCDM) approach aimed at sorting those failure modes potentially involving critical components into risk classes for interventions prioritisation and maintenance control. In this context, the sorting technique ELimination Et Choix Traduisant la REalité (ELECTRE) TRI is applied to support in finding the root causes that can be eliminated for failure prevention and/or minimization. This methodology presents the advantage to not rely on comparisons (as well as on their transitivity) between pairs of elements, simplifying computations for complex systems. To be sorted, decision elements are indeed compared with single reference profiles and the final assignment may constitute a valid alternative to the traditional ranking of failures achievable by other MCDM techniques and, among others, consistency-based methodologies. The proposed approach will be eventually applied to a case study from the industrial reality.