AIM-FLIX – AI-driven inventory management for flexible, and efficient nextgen supply chains

Traditional inventory management approaches, which are based on static rules, fixed safety stocks, or independent forecasts across the supply chain echelons, often fail to anticipate these dynamics. This leads to emergency shipments, excess stock, high returns, and increased environmental impact.

The AIM-FLIX project is building on the results from the AIM-TRUE project, which developed novel AI/ML tools and demonstrated the feasibility of meta-learning for spare parts demand forecasting in realistic scenarios and at a global scale. The AIM-FLIX project takes the next step towards Reinforcement Learning (RL)–driven multi-echelon policy optimization across the entire interconnected supply network. Our goal is to enable adaptive, network-wide policies for inventory planning and rebalancing, dynamically coordinating service levels, costs, and sustainability considerations. The project will explore solutions such as shared stocking (treating inventories across warehouses and dealers as a shared pool), and survival analysis (understanding the probability of adverse events, such as stock-outs, returns, or obsolescence, occurring over time)—all to improve responsiveness, reduce duplication, and optimize stock placement. Measurable targets include reducing rush orders by 20% and parts scraping by 15%.

The project will be conducted over three years, starting from January 2026. Project partners are Volvo Service Market Logistics (SML) as the problem owner and the leading actor able to implement more resource-efficient and sustainable logistical solutions, Rejmes Transportfordon (RT) representing the dealer network as the crucial entity affected by the proposed changes, and the Center for Applied Intelligent Systems Research at Halmstad University (CAISR) as a research environment and project coordinator.