Supply Chain Inventory Control-Methods for Coordination and Evaluation

Abstract: Thanks to the fierce competition and the globalization of markets as well as compa¬nies, supply chain management has emerged as one of the most urgent manage¬rial issues in business today. This thesis is devoted to one aspect of this vast problem area - the coordinated management/control of supply chain inventories. The objective is: To develop methods and models for coordinated inventory control in divergent supply chains with stochastic demand, subject to different control structures and information availability. Methodologically the research belongs to the field of operations research and sto¬chastic multi-echelon inventory theory, and is presented in the form of six scien¬tific articles preceded by a summa¬rizing introduction. One common feature of the six papers is that they all focus on models for inven¬tory control in generic supply chains con¬sisting of one central warehouse serving an arbitrary number of retailers experi¬encing stochastic demand. Otherwise, the different models display a diversity of dif¬ferent in¬formation and control struc¬tures. Paper I presents a new efficient method for determining near optimal reorder points in highly de¬centralized supply chains with batch ordering. The suggested coordination procedure can be inter¬preted as a game or a negotiation process. It successively solves a number of single stage problems, using a derived penalty cost at the warehouse to achieve coordination. Paper II generalizes the model in Paper I to systems with com¬pletely non-identical retailers. In Papers III, IV and V, the goal is to design new replenishment and allocation poli¬cies aimed at more efficient use of conventional inventory and demand information. Paper III presents a new ware¬house replenishment policy for central¬ized control in continuous review systems with batch ordering. An exact cost evaluation technique is de¬rived, and numerical results indicate that significant cost sav¬ings can be achieved by using the new policy. Paper IV investigates new, computationally tractable heuris¬tics for cen¬tralized control in periodic review systems. The high potential cost sav¬ings, par¬ticu¬larly in case of large differences in backorder costs and demand variabil¬ity across retailers, are illus¬trated in a numerical study. Paper V also deals with peri¬odic review systems, focusing on how to suc¬cessively allocate stock to the retailers over the ware¬house order cycle. The main result is the deriva¬tion of a new type of policy, asymp¬toti¬cally optimal when the number of retailers approaches infinity. Based on the asymptotic analysis, a number of heuristics for supply chains with finitely many retailers are presented and evaluated numerically. Both centralized and decentralized settings are considered. Paper VI, finally, considers a situation with extended customer demand information, where each customer order entails a future delivery date. Exact and approximate methods for cost evaluation un¬der two different allocation strategies in a continuous review environment are presented. The large potential benefits of fully using this type of advance-order information are illustrated in a nu¬merical study.