The Cost of Quality in Circular Economy: Balancing AI Investment and Profitability in CRD Waste Management

The management of Construction, Renovation, and Demolition (CRD) waste faces challenges due to low material sortability and limited effectiveness of manual sorting, which lead to impurity rates and financial losses. This study proposes the integration of Industry 4.0 technologies, AI-driven computer vision and robotic sorting into a Reverse Logistics Network Design (RLND) under uncertainty. We develop a two-stage stochastic Mixed-Integer Quadratically Constrained Program (MIQCP) to model the nonlinear relationship between investment and sorting efficiency, using a cost of quality function to reflect diminishing returns. To address the non-convexity of the model, a McCormick linearization approach is employed with Gurobi optimizer. Results from a case study in Quebec demonstrate that AI-enabled sorting improves effective recycling rates from a baseline of 0.47-0.58 to 0.72-0.85. However, this transition results in an 11.1% reduction in short-term profitability. These results illustrate the cost of quality required to guarantee the high-purity recycled materials necessary for a sustainable circular economy.