A signature-based approach for reliability modeling and maintenance optimization of a coherent system

In reliability context, mean residual lifetime (MRL) models and proportional hazard (PH) models are widely used for reliability analysis and failure prediction. Since maintenance actions typically are based on aforementioned reliability metrics, recent researches have been focused on the extensions of these two models for accurate decision-making. In this paper, using signature arguments, we propose a state-dependent mean residual life (SDMRL) model and demonstrate its applicability in both reliability and maintenance field. In preference to existing models, the SDMRL modeling can be applied to coherent systems and account both changes in the operating environment and for the natural ageing process of the system. Assuming a bivariate threshold-type policy, the paper aims at minimizing the long-run average maintenance cost per unit time by determining appropriate inspection intervals and optimal replacement policy. The proposed model and the behavior of the optimal solutions as the model parameters vary are illustrated through numerical examples.