A mixed integer linear programming model representing a two-echelon manufacturing system was implemented. Optimal decisions could be made about raw material/component provisioning by using the model. The model was programmed by using an algebraic modeler which was then integrated into a computational tool from which defining parameters could be managed as well as consulting the results once the model had been executed. The model was validated on a real manufacturing system; besides providing good representation of the system, optimal provisioning decisions were also reached. The article emphasises that such decisions cannot be made by using usual MRP reasoning.