摘要：Production control problem for failure-prone systems that utilize both manufacturing and remanufacturing in production processes and are subject to return uncertainty and variability is studied. Proposed feedback control policies are based on the imprecise measurements. The return flow is unknown and its rate is composed of seasonal and random components. To cope with measurement errors and return variability, estimation and forecasting modules are designed and incorporated into a control loop. Market demand is supposed to be constant although may contain an unknown component (much smaller than the return). Kalman filter technique is used for estimating inventory levels and return rate online from noisy inventory measurements. Obtained estimates are used in production control procedure ensuring that the policy continuously adapts to current market conditions. The system behavior under large return rate variations is studied and cost effectiveness of the proposed policies is demonstrated.
