A data-driven and risk-based prudential approach to validate the DDMRP planning and control system

In this paper, we study the single-item dynamic lot-sizing problem in an environment characterizedby stochastic demand and lead times. A recent heuristic called Demand Driven MRP,widely implemented using modern ERP systems, proposes an algorithm that is will effectivelytackle this problem. Our primary goal is to propose a theoretical foundation for such a heuristicapproach. To this aim, we develop an optimization model inspired by the main principlesbehind the heuristic algorithm. Specifically, controls are of the type (s(t); S(t)) with time-varyingthresholds that react to short-run real orders; in this respect, control is risk-based anddata-driven. We also consider service levels derived as tail risk measures to ensure fulfillmentof realized demand with a predetermined probability; in this respect, our approach is prudential. Finally, we use our model as a benchmark to theoretically validate and contextualize theaforementioned heuristic