AI-Driven Circular Logistics: Integrating Reverse Supply Chains for Sustainability

Abstract

The transition toward circular-economy principles has increased the need for efficient and sustainable logistics systems, yet reverse logistics remains characterized by uncertainty, fragmentation, and operational inefficiencies. This study proposes and evaluates a simulation-based AI-driven circular logistics framework that integrates machine-learning predictive analytics with multi-objective optimization to support reverse logistics decision-making under uncertainty. Using a medium-scale simulated supply chain environment, the framework is evaluated through comparative scenario analysis involving Baseline, AI-Forward, and AI-Circular logistics configurations. The simulation results suggest that integrating predictive analytics with circular logistics strategies may improve economic and environmental performance by reducing logistics costs and carbon emissions while enhancing resource recovery. Sensitivity analysis indicates that the framework remains robust under varying return rates, demand conditions, forecasting errors, and objective-weight configurations within the modeled environment. The study contributes an integrated conceptual and analytical framework linking artificial intelligence, circular economy principles, and supply chain management and provides a foundation for future empirical validation of AI-enabled circular logistics systems. The findings should be interpreted as illustrative and conditional on the assumptions of the simulation environment rather than as evidence of universally achievable operational outcomes.