An Approximation Scheme for a Class of Operations Management Problem with an O(TK) Performance Bound
DescriptionFacing uncertain demand and fierce market competition, company managers in retailing and manufacturing industries need to better manage their inventory to reduce cost, and also increase their revenue by making smart pricing decision. Although inventory control or joint pricing and inventory control is a main research stream that has received considerable attentions in operations management literature, great challenges still exist in some important practical applications. This is especially true for those periodic review problems where the objective is to optimize the total cost/profit over a planning horizon. For this class of problems, the optimal policies are usually too complicated to be implemented in practice. In this project, we solve this difficulty by proposing a close-to-optimal and easy-to-implement heuristic policy. By developing a new mathematical tool called weak K-convexity, we can establish a worst-case bound on the performance of the heuristic policy, which significantly improves current result in the literature. Meanwhile, in the case of incomplete demand information where a retailer has only exercised very few price points, the concept of weak K-convexity can provide us some guidance on how to learn the demand as a function of price. As this information is crucial for making the pricing decision, we develop a demand learning procedure basing on weak K-convexity. Overall, this project uses a new methodology called weak K-convexity to advance the applications of academic work into practice by removing certain barriers between them.
|Effective start/end date||1/07/21 → …|