Optimization of Reverse Supply Chain For End-of-life Products

Abstract

The evolution of circular economy has led to the adoption of circular supply chains, allowing original equipment manufacturers to recycle, reuse, and remanufacture end-of-life (EoL) products containing valuable resources. Efficient reverse supply chain (RSC) management becomes crucial, optimizing logistics and information flow for these products to enhance resource utilization, minimize waste, and achieve circular supply chains. The foundation of an effective RSC management lies in a well-structured reverse logistics network. Accurately predicting EoL products quantities is essential due to limited samples in the recycling industry’s early stages and the absence of standard regulations in certain regions, which challenge the precision of existing prediction approaches. A critical issue of reverse logistics network (RLN) design is optimizing facility locations, yet few studies address this with a triple bottom line perspective. Design challenges emerge when addressing indeterminate phenomena, particularly uncertain parameters within complex reverse flow. Current approaches have limitations, notably the insufficiency of observational data in managing EoL products. Additionally, existing heuristic algorithms may face difficulties in accurately capturing inter-relational constraints among multiple objectives. Furthermore, RSC management faces issues of information transparency and consistency, which hinder stakeholder decision-making and the expansion of the circular supply chain. To fill these gaps, this research proposes a Stacking-based ensemble model integrated with data augmentation method, enhancing prediction accuracy for the quantity of EoL products. Then, a multi-objective mathematical model, considering economic, environmental, and social objectives, is introduced for optimizing facility locations in RLN design. A method integrates uncertain multi- objective programming and NSGA-III to solve the mathematical model. Finally, to improve information transparency and traceability and collaborative practices within the RSC, this research suggests a conceptual framework for managing RSC information, incorporating digital product passports via blockchain technology.