Optimal Component Procurement Sequence in Assembly Systems under Supply Uncertainty

Abstract

We study an assembly system for a final product consisting of multiple components. Each component is sourced from a supplier with random production yield. We evaluate the benefits of postponing the procurement decisions of some components until observing the yield realizations of the other components. We develop an analytical framework to determine the optimal component procurement sequence based on two related performance measures---threshold price (i.e., the minimum market price that ensures a positive profit for the assembler) and profit. For a two-component assembly system with uniform yield distributions, we show that the procurement sequences generated based on the two performance measures are largely consistent. The assembler should postpone the procurement decision of the component with a less variable yield and/or a higher unit effective cost. We extend our analysis to general yield distributions and to general n-component assembly systems, and provide an efficient heuristic for determining the optimal procurement sequence. In addition, we further quantify the benefits of procurement postponement as an operational strategy to mitigate the adverse impact of yield uncertainty, and show that the benefits become larger when the unit component costs are higher. Finally, a comprehensive set of numerical studies is conducted to demonstrate the robustness of our findings.

Keywords: Random yield, procurement postponement, assembly system, threshold price

JEL Classification: C02, D2, L6, M11