A general maintenance policy for a multi-state deterioration system with multiple choices of imperfect maintenances

This study proposes a general maintenance policy for a system whose deterioration process is in a Gamma process. Such systems under scheduled inspection will be modeled by a multi-state Markov system with time-dependent transition probabilities. Such a policy is mainly to determine, at each inspection time nd, the optimal maintenance action from multiple candidates under the conditions that: (1) the transition probabilities and the risks of each maintenance action are updated at each nd, (2) both the times spent on inspection and maintenance are ignored, etc. Finally, a numerical example is presented through simulation to illustrate how such a policy can be carried out.