Training and repair policies for stand-by systems

This research is concerned with developing repair and training strategies for stand-by equipment which maximise the time until the equipment is unable to respond when it is needed. Equipment can only be used if it is in an operable state and the users have had sufficient recent training on it. Thus it is necessary to decide when to maintain/repair the equipment and when to use the equipment for training. Both actions mean the equipment is not readily available for use in an emergency. We develop discrete time Markov decision process formulations of this problem in order to investigate the form of the optimal policies which maximise the expected survival time until a catastrophic event when an emergency occurs and the equipment cannot respond. We also calculate the solution in a number of numerical examples.