News-boy problem with unreliable suppliers

Abstract

In this paper, we investigate the single-period inventory control problem with an unreliable supplier while facing stochastic demand. We focus on two cases: when the distributions of demand and random yield are known and when only information of mean and variance are available. We derive the mathematical formulation for both cases, where maximum entropy and distribution-robust frameworks are considered. Then, we study a dual sourcing model with one perfectly reliable supplier and one unreliable supplier. Optimal sourcing strategies are characterized, and findings and insights are discussed. Lastly, we compare the effectiveness of the distribution-robust approach, maximum entropy approach and sample average approximation approach. It is observed that, in general, the maximum entropy approach outperforms the other two alternatives.