Continuous-time Optimal Control of Deteriorating Inventory/Production Systems Using a Demand Observer

Abstract A recently developed technique in motion control is applied to solve the demand variation problem in an inventory/production system where items deteriorate at unknown constant rate. A feedback control technique called demand observer (DEMO) is applied to the inventory-level dynamics to force it to behave as a well-defined nominal model. As a result, any production policy based on linear control theory used in conjunction with the demand observer will be robust to changes in the demand and uncertainty in the deteriorating rate. In addition, the technique works for any moderately varying demand and for finite or infinite planning horizons. To illustrate the proposed production control structure, a standard linear optimal control scheme is used with a simple version of the demand observer. Finally, the overall inventory/production control is tested for different demand patterns. The simulation results showed the high performance robustness of the proposed structure for large demand variations and deteriorating rate uncertainty.