Robust reverse production system design for carpet recycling

Abstract

It is estimated that complete carpet recycling would avoid an estimated US annual landfill cost of $65 million, while simultaneously recovering lost material valued at $750 million. Designing an adequate reverse production system is critical to the economic viability of recovering this lost value. We develop a robust-mixed-integer linear programming model to support decision-making for reverse production infrastructure design. Our robust model seeks solutions close to the mathematically optimal solutions for a set of alternative scenarios identified by a decision-maker. To demonstrate the approach, a representative industrial case study is given for a large-scale carpet recycling problem. A robust solution is found that appraises the impact of two major sources of uncertainty, volumes of carpet collected and price of a key recycled material.