Distributionally Robust Newsvendor with Moment Constraints

💡 Research Summary

The paper extends the classical single‑period newsvendor problem by incorporating both moment information (mean μ and variance σ²) and distributional ambiguity measured with the Wasserstein distance. Starting from the empirical distribution Pₙ built from observed demand data, the authors define an ambiguity set U_δ(Pₙ) consisting of all probability measures whose Wasserstein distance to Pₙ does not exceed a radius δ. In addition, any admissible demand distribution must satisfy the moment constraints M₁(X)=μ and M₂(X)=μ²+σ². The decision maker seeks an order quantity Q ≥ 0 that minimizes the worst‑case expected cost
c₁ E