The impact of demand aggregation through delayed component allocation in an assemble-to-order system

We consider an assemble-to-order system in which multiple products are assembled from a common component and a set of product-dedicated components. Component capacities are chosen prior to a finitehorizon selling season, and the common component is allocated to the products based on observed demands. We propose a collection of allocation mechanisms involving varying degrees of demand aggregation, ranging from a scheme under which all demands are observed prior to making the allocation decision to allocations made for each arriving demand. In this context, we explore the impact of the allocation scheme on sales, profits, and capacity decisions, including the degree of capacity imbalance. We find that the benefit from increased demand aggregation is closely linked to the degree of capacity imbalance: profit gains from delayed allocation tend to be higher in systems in which the optimal capacity portfolio is highly unbalanced when the allocation decision is made after observing all demands. We develop insights into what detailed system parameters lead to the largest gains from demand aggregation and also explore the trade-offs associated with the choice of an allocation scheme when customers exhibit impatience if the allocation scheme forces them to wait to be served.