Adaptive Inventory Management in Global Supply Chains Using Digital Twins and Reinforcement Learning for Disruption Resilience and Sustainability

Abstract

Global supply chains are increasingly vulnerable to disruptions arising from geopolitical instability, climate-related events, and volatile consumer demand. Traditional inventory management systems, largely reliant on static forecasts and rigid safety stock parameters, struggle to cope with this complexity and lack adaptability in the face of real-time uncertainties. This has spurred the emergence of digitally enhanced, intelligent inventory systems that prioritize both disruption resilience and sustainability. This article proposes an adaptive inventory management framework that integrates Digital Twin technology with Reinforcement Learning (RL) to enable real-time, feedback-driven optimization across global supply networks. Digital Twins virtual replicas of physical supply chain entities serve as dynamic environments where RL agents continuously simulate, learn, and optimize inventory strategies based on changing external and internal conditions. The model incorporates multi-dimensional input streams such as transportation delays, supplier reliability indices, carbon footprint data, and demand fluctuations to inform sustainable and resilient inventory decisions. By embedding this system into global logistics networks, inventory policies become more responsive to unanticipated events such as port closures, raw material shortages, or energy price shocks. Simulated case studies involving manufacturing and retail supply chains demonstrate that the Digital Twin-RL hybrid system significantly outperforms conventional models in terms of stock availability, cost efficiency, and emissions reduction. Moreover, it supports predictive anomaly detection and prescriptive decision-making to reduce waste, improve lead-time reliability, and promote environmental accountability. This research underscores a critical paradigm shift in inventory management one that blends data-driven intelligence with digital simulation to achieve adaptive, resilient, and sustainable supply chain operations in a volatile global environment.