A policy for the maintenance of a multistate multicomponent series system

In real life situations various levels of performance for both systems and components are identified. Deterioration of systems and components has been considered by many maintenance and replacement models in the literature, but most of these models are for single component. Our study concerns a multistate multicomponent maintenance model. We extended the model studied by Van der Duyn Schouten and Vanneste (1993) with four states, (0,1,2,3), to a multistate system with 5 + 1 states, (0, 1,..., S). While classifying these S +1 states, we use two different classification schemes corresponding to different real life situations. Then, we analyze our model under both exponential and Erlang sojourn time distributions for both state classification schemes. As a result of our analysis, we derive expressions for the long-run average cost per unit time. We used the randomization method to evaluate the transient state probabilities during the implementation of the average cost expressions.