Deterministic allocation models for multi-period demand fulfillment in multi-stage customer hierarchies

In manufacturing environments with heterogeneous (e.g., because of different profit margins) customers the optimal matching of available supply with dynamic demand is a challenging task as soon as supply becomes scarce. In Make-to-Stock systems, Demand Fulfillment first allocates these scarce resources (in form of Available-to-Promise - ATP) to customer segments on basis of their forecasted demand. The resulting quotas are then consumed (“promised”) when real customer orders arrive. In a multi-stage sales hierarchy, this allocation process often has to be executed level by level, on basis of decentral, aggregate information only. Decentral, multi-period, deterministic linear and non-linear programming models are proposed approximating the first-best benchmark of a central, multi-period allocation planning with full information. Roll over simulations have been performed to obtain insights about the behavior of the proposed decentral models and to demonstrate their benefits in comparison with common quantity-based methods, especially for high levels of customer heterogeneity and high shortage rates.