Control of production processes subject to random shocks

Quality of industrial production processes may suffer from wear-out phenomena continuously growing in time, and from sudden disturbances occurring at random time. Either of these affects process quality and thus process yield. Consequently, there should be counter-actions, which aim to remove signs of wear before they result in a decrease of process yield on the one hand, and which detect disturbances before their impact has become serious. In this paper, it is assumed that the wear-out phenomena are sufficiently compensated by means of continuous maintenance. It remains to determine an appropriate monitoring policy in order to detect disturbances. A unified economic approach is proposed and investigated, including different modes of process monitoring together with a decision function reacting simultaneously on shifts in the process mean and process variability.