Practice-inspired contributions to inventory theory

The research field of inventory control has a clear practical aim: managing real inventories of real companies, to achieve the optimal trade-off between availability and costs. Yet, there are many discrepancies between inventory control in theory and in practice, which leads to worse outcomes than were to be expected theoretically. Motivated by observations from practice, this thesis adapts and extends existing inventory control models. The main contribution of this thesis focuses on the generally missing link between inventory control and demand forecasting. Inventory control models assume that the probability distribution of future demand and all its parameters are known completely. In practice, demand has to be forecasted, and this forecast will inevitably have an error. This thesis contributes to both better demand forecasting and better integration of demand forecasts in the inventory control models. The other contributions of this thesis are to the inventory control models themselves. Inventory control models can be based on continuous or periodic reviews of the inventory level. It is common in the literature to let replenishments coincide with these review moments. We show that if this restriction is relaxed, then it is typically optimal to not replenish at or around the review moments. Finally, we study the stocking problem of a repair service engineer and take into account that spare parts may have to be shipped from elsewhere before they can be restocked in the engineer’s kit.