Distribution-robust single-period inventory control problem with multiple unreliable suppliers

This paper presents a distribution-robust approach to the single-period inventory control problem with multiple unreliable suppliers and stochastic demand. Using the mean and covariance that are assumed to be the only known data, the proposed approach tries to find a solution that performs well independent of the specific distributions. The approach can provide a way to effectively hedge against risks coming not just from uncertainty itself but also from inaccurate estimation. Moreover, the resulting optimization model turns out to be computationally tractable and amenable to mathematical analysis. Using the model, we provide a complete characterization of the optimal sourcing strategies when there are two suppliers: a perfectly reliable supplier and an unreliable supplier. Various interesting issues relevant to sourcing problems, including the value of supplier diversification and yield quality improvement, also are addressed.