Integrating Reinforcement Learning and Generative AI for Dynamic Inventory Rebalancing and Demand-Driven Replenishment in Multi-Echelon Supply Chains

This study investigates the integration of reinforcement learning (RL) and generative artificial intelligence (GenAI) to optimize dynamic inventory rebalancing and demand-driven replenishment across multi-echelon supply chains. By leveraging GenAI to generate synthetic demand scenarios and RL to adaptively manage inventory flows, the proposed hybrid model addresses the complexities of decentralized decision-making, demand volatility, and operational inefficiencies. A modular architecture is developed, combining cloud-native simulation, interpretability mechanisms, and fairness auditing to ensure transparency, ethical compliance, and adaptability. Experimental results reveal significant improvements in stockout rates, turnover efficiency, and cost reduction compared to conventional models. The system also demonstrates strong resilience under disruption scenarios and aligns with ethical AI deployment frameworks championed by leading scholars. This research offers a scalable, data-driven solution for real-time supply chain optimization, contributing to the broader discourse on intelligent logistics automation and responsible AI adoption.