Roughly 4 billion pounds of carpet are disposed of in the United States eac
h year. This carpet is composed of a significant fraction of nylon, polypro
pylene, and polyester fiber. A key limiting factor to recycling is effectiv
e design and development of the reverse production system to collect and re
process this large volume of valuable material. A reverse production system
is composed of material and chemical recycling functional elements interco
nnected by transportation steps. In this article, we develop a mixed-intege
r programming model to support decision-making in reverse production system
design. To illustrate its use and applicability, we apply the model to a r
epresentative U.S, carpet recycling industrial case study. The overall econ
omic feasibility of recycling is strongly dependent on the volumes that can
be expected from investments in collection infrastructure. The geographic
location of processing centers influences the network economics, and the su
bdivision of recycling tasks to avoid the shipment of low value material is
proposed as an effective strategy for carpet recycling.