THESIS
2006
xiii, 94 leaves : ill. ; 30 cm
Abstract
Rolling-horizon planning is a natural planning method widely used in the manu-facturing and distribution sectors. However, while the effectiveness of the stan-dard rolling-horizon scheduling has been examined and confirmed, we have not found any recent study that develops a rolling-horizon replenishment policy incor-porating Quantity flexibility features and investigates its effectiveness in different operational environments. Facing this research opportunity, we first propose a rolling-horizon replenishment modeling framework under which a buyer can up-date demand information and inventory status, modify order quantities commit-ted previously for periods beyond the frozen fence, and place an advanced order for a new period at the end of the rolling-horizon. Using a finite-horizon stoch...[
Read more ]
Rolling-horizon planning is a natural planning method widely used in the manu-facturing and distribution sectors. However, while the effectiveness of the stan-dard rolling-horizon scheduling has been examined and confirmed, we have not found any recent study that develops a rolling-horizon replenishment policy incor-porating Quantity flexibility features and investigates its effectiveness in different operational environments. Facing this research opportunity, we first propose a rolling-horizon replenishment modeling framework under which a buyer can up-date demand information and inventory status, modify order quantities commit-ted previously for periods beyond the frozen fence, and place an advanced order for a new period at the end of the rolling-horizon. Using a finite-horizon stochas-tic dynamic programming model, we show that the optimal ordering policy is a dual-threshold plus base-stock type for models with a three-period rolling hori-zon. To compute the thresholds and base-stock level, we provide exact formulas for the model with three-period demand forecasts and design efficient algorithms for models with more than three-period forecasts. An important and challenging problem for finite-horizon inventory control is the forecast horizon. We construct two monotone bounds and develop an algorithm to obtain the forecast horizon for our problem exactly. Second, when demands are correlated during the dy-namic forecasts, we give the optimality condition for the myopic policy. Then, we develop bounds for optimal ordering quantities and cost error bounds to assess the efficiency of heuristic policies. Finally, we consider the design of coordination mechanism for a two-stage decentralized supply chain in which a supplier provides a single product to a buyer who faces the external market. On the one hand, for the decentralized chain, we explore a wholesale price plus price protection mech-anism to achieve channel coordination. On the other hand, we consider how to set contract parameters to maximize the supplier's own expected revenue.
Post a Comment