Maintenance optimization for dependent two-component degrading systems subject to imperfect repair

Appropriate maintenance policies play an important role in improving system availability and ensuring safe operation. Seeking optimal maintenance policies for technical systems has been widely pursued by reliability engineers and researchers. In this paper, we propose a maintenance optimization method that is applicable to dependent two-component systems subject to degradation and imperfect repair. We consider both economic and stochastic dependencies between the components and establish a random-effect imperfect repair model to realistically model the degradation process and maintainability of components. Moreover, we model the maintenance problem under the infinite horizon using the Markov decision process and obtain the optimal solution via value iteration algorithm. Structural insights are gleaned using the stochastic orders. A numerical example is then presented to illustrate the proposed methods. We discover that the characteristics of imperfect repair can considerably influence the optimal policies. Specifically, the mean effect of imperfect repair has a larger influence on maintenance decisions while the influence of imperfect repair variability effect is relatively small.