Abstract
In most previous studies of maintenance decision-making, systems were assumed to deteriorate in accordance with a stationary state transition law. However, the deterioration of a system sometimes depends on its age (calendar time, total operating time, etc). In this research, optimal maintenance decision-making was investigated for systems with age-dependent deterioration. The problem was formulated as a non-stationary Markov decision process. The properties of the optimal expected cost function were examined, and the optimal maintenance decision-making policy was found to be given by a two-dimensional control limit policy with respect to both the system's deterioration state and age under certain conditions. A numerical example of a power transformer demonstrated that a control limit policy that takes both state and age into consideration is optimal for this system and is also better than the existing maintenance guideline. Such an age-dependent maintenance policy supports the implementation of a system-specific maintenance plan.
