Optimal inventory-production control in discrete manufacturing systems

The problem of inventory-production control in discrete manufacturing systems in the presence of stochastic disturbances and measurement noise is considered. The proposed control algorithms are based on the internal model control principle. In such a case the closed-loop system has a selective invariance properties and the influence of stochastic disturbances may be sufficiently decreased. The problems of delay modeling, identification and compensation in inventory control systems are also considered. The generalized predictor-based delay compensating controller as well as design procedure is proposed and comparison with known methods has been performed. The new type of random delay model is developed using Hidden Markov Model technique. It is shown that such an approach leads to the simple varying delay detection and random delay identification algorithms based on ordinary recurrent estimation technique. Copyright (C) 2000 IFAC.