An optimal maintenance policy for deep submergence equipment

This report presents an optimal maintainance policy for deep submergence equipment for which the following assumptions hold: (i) The system deteriorates with time and finally fails to operate; (ii) It is impossible either to repair the system or to observe its level of deterioration without bringing it to the surface; (iii) Even in a badly deteriorated state the system can perform its function although with a substantial loss; (iv) The observation and repair of the system are quite expensive because of assumption (ii). Using a Markov decision process as a model, maintainance policy is derived which is optimal in the sense that it maximizes the return from the operation of the system with respect to the high costs for its observation and repair.