LP-based heuristics for the Capacitated Lot Sizing Problem

Master Production Scheduling is a cornerstone of MRP II systems, since it represents the link between marketing and production. To overcome the limitations of MRP Il systems, we aim at finding a good and realistic MPS by optimization techniques, rather than by a trial-and-error manual procedure. Due to the complexity of the resulting mathematical model, heuristic solution strategies must be devised. In this paper we report computational experiments on the application of approximate solution methods, based on Linear Programming, to a simplified version of the problem, namely the Capacitated Lot-Sizing Problem (CLSP). We show the usefulness of suitable model reformulations and of interior point solution methods.