| Customers and downstream supply chain partners often place, or can be induced to place, orders in advance of future requirements. We show how to optimally incorporate advance demand information into periodic review inventory models. For single stage models we show that state dependent base stock and (s, S) policies are optimal. The state space is given by a modified inventory position (that nets known requirements over the protection period), and observed demands beyond the protection period (the lead time plus a review period). A horizon result is shown to prevail when observed demands after the protection period are sufficiently high. In this case it is optimal to ignore observed demands beyond the protection period. We show that myopic policies are optimal if setup costs are zero and costs and demands are stationary. Most results extend to multi-stage serial systems. While the state space for serial systems appears to be formidably large, we show how to decompose the problem across stages and how to reduce the state space at each stage via modified echelon inventory positions. In addition, we show that optimal policies are very easy to compute when costs and demands are stationary. Unlike serial systems optimality results do not extend to one-warehouse multi-retailer distribution systems. We develop, however, a lower bound (based on a relaxation of the exact formulation) and a close-to-optimal heuristic that incorporates advance demand information. With the aid of extensive numerical study we obtain additional insights about the optimal control of inventories in the presence of advance demand information. We also provide managerial insights about the value of advance demand information and point out cases where additional information has no operational value. |
