Abstract: By postponing the final assembly of products until the actual demand is observed,ATO (Assemble-to-Order) systems can reduce the over-production cost and risk that usually exist in the MTS (Make-to-Stock) systems. However,when the future assembly capacity is constrained,the assembler must assemble a certain quantity of final products in advance so as to maximize his revenue. This paper considers a single-period ATO system that faces both uncertain future demand and uncertain assembly capacity. Before the actual demand is observed,the assembler needs to procure all the necessary components,and pre-assemble some final products. After the actual demand and the actual available capacity are observed,he needs to assemble the rest of the product (if needed),so as to satisfy the customer as much as possible and thus maximize his profit. Based on a two-stage optimization model,the optimal component-inventory and optimal production decisions of the assembler are studied in this paper.We further extend the model to a more complicated situation where emergent outsourcing can be acquired,at an extra cost,to increase the assembly capacity. Our research results provide some interesting managerial insights for the scheduling of ATO systems in an uncertain environment.

Key words: assemble-to-order systems, inventory optimization, stochastic demand, uncertain capacity, assembly advance