Combining MPC and LD Analysis in Supply Chain Inventory Control Problem

The present paper approaches the inventory control problem in supply chains. A classical solution of this problem is based on base-stock policy which, more recently, has been improved using large deviations (LD) techniques and introducing the concept of service level constraint. Base-stock policies are widely used in real-world organizations and investigated by academia, however the solution provided by these policies are non-smooth and leads to high set-up costs. Our idea is to use a model predictive control (MPC) technique to mimic the base-stock policy using the same large deviations analysis of the system but achieving both lower inventory and production costs, without violating the imposed service level constraint