5. Data-Driven Inventory Management

Over the last decades that were characterized by just-in-time supply chains, inventories were highly unpopular and avoided wherever possible in order to reduce capital lockup to a minimum. However, severe supply disruptions and inflation risk in the early 2020s showed that inventories can be of strategic importance in times of high volatility and economic uncertainty. This chapter focuses on single-item inventory optimization and addresses both deterministic and stochastic models as well as single-period and multi-period approaches to optimally control stock levels in a data-driven manner. The chapter introduces important inventory metrics for performance management, analytical safety stock planning models and latest developments in machine learning and deep learning for inventory management applications.