Manufacturing network configuration in supply chains with product recovery

Efficient implementation of product recovery requires appropriate network structures. In this paper, we study the network design problem of a firm that manufactures new products and remanufactures returned products in its facilities. We examine the capacity decisions and expected performance of two alternative manufacturing network configurations when demand and return flows are both uncertain. Concerning the market structure, we further distinguish between the case where newly manufactured and remanufactured products are sold on the same market and the case where recovered products have to be sold on a secondary market. We consider network structures where manufacturing and remanufacturing are both conducted in common plants as well as structures that pool all remanufacturing activities in a separate plant. The underlying decision problems are formulated as two-stage stochastic programs with recourse. Based on numerical studies with normally distributed demands and returns, we show that particularly network size, investment costs of (re-)manufacturing capacity, and market structure have a strong impact on the choice of a network configuration. Concerning the general role of manufacturing configuration in a system with product recovery, our results indicate that the investigated structures can lead to very different expected profits. We also examine the sensitivity of network performance to changes in return volumes, return variability and correlation between return and demand. Based on these results, we find that integrated plants are more beneficial in the common market setting. This relative advantage tends to diminish when demand is segmented, thus investing in more specialized, dedicated resources should be considered.