Theorem 13 — Gelbrich hull
DisprovedWassersteinDRO.Gelbrich.gelbrich_hull_containmentoperations-researchprobabilityrobust-optimization
If the nominal distribution has mean vector and covariance matrix , then for every : the Gelbrich hull is an outer approximation of every type- Wasserstein ball with , using only first- and second-order moment information about the nominal distribution.
Preamble
import Mathlib import Definitions.Def_WassersteinDRO_Gelbrich_ambiguitySet import Definitions.Def_WassersteinDRO_Gelbrich_gelbrichHull open MeasureTheory
Formal statement
namespace WassersteinDRO.Gelbrich
/-- Theorem 13 (Gelbrich hull), Kuhn et al. 2019, p. 18 — the goal theorem: if the nominal
distribution `P̂N` has mean vector `μ̂` and covariance matrix `Ŝ`, then the Wasserstein
ambiguity set is contained in the Gelbrich hull, `B_{ε,p}(P̂N) ⊆ G_ε(μ̂,Ŝ)`, for every `p ≥ 2`.
This is the outer approximation that Corollary 1 turns into a tractable upper bound on the
worst-case risk, and that Theorem 16 (not formalized in this chunk — see `STATUS.md`) sharpens
to an equality for quadratic losses and elliptical nominal distributions. -/
theorem gelbrich_hull_containment {m : ℕ} (ε p : ℝ) (hp : 2 ≤ p)
(Ξ : Set (EuclideanSpace ℝ (Fin m))) (PN : Measure (EuclideanSpace ℝ (Fin m)))
(μhat : EuclideanSpace ℝ (Fin m)) (SigmaHat : Matrix (Fin m) (Fin m) ℝ)
(hμ : meanVector PN = μhat) (hS : covarianceMatrix PN = SigmaHat) :
ambiguitySet ε p Ξ PN ⊆ gelbrichHull ε Ξ μhat SigmaHat := by sorry
end WassersteinDRO.Gelbrich
Source
Kuhn, Mohajerin Esfahani, Nguyen & Shafieezadeh-Abadeh, Wasserstein Distributionally Robust Optimization, INFORMS TutORials 2019, Theorem 13, p. 18
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.