Stochastic lot-sizing for a multi-echelon remanufacturing system
Résumé
We consider a remanufacturing system involving three production echelons: disassembly of returned products, refurbishing of recovered parts and reassembly of remanufactured products. This leads to the formulation of a three-echelon lot-sizing problem with a general product structure, product returns and lost sales. We first consider a deterministic variant of the problem and focus on its resolution through Cut & Branch and Branch & Cut algorithms. We then investigate a stochastic variant of the problem in which the uncertainties on the quantity of returned products and on the demand for remanufactured products are explicitly taken into account. We propose a multi-stage stochastic programming approach relying on a scenario tree to represent the uncertainty. The extension of the algorithms proposed for the deterministic problem is then discussed. Preliminary computational results are presented.