§5.2, proof of Proposition 3, p. 20 — (19) has an optimal solution satisfying (19b) with equality
ProvedMomentDRO.WorstCov.exists_optimal_tightloewner-orderoptimal-solutionp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1semidefinite-programming
Let , , , and . The SDP (19) of p. 19 (with objective ) has an optimal solution for which constraint (19b) holds with equality:
That is, is feasible, no feasible point has a larger objective value, and the sum of the matrices fills the whole budget .
This is the step where the largest covariance enters: the tight matrix constraint becomes the second moment of the worst-case distribution built from .
Formalization Note The page starts from "an optimal assignment "; the statement includes the existence of an optimum, which the page uses implicitly.
Preamble
import Mathlib import Definitions.Def_MomentDRO_WorstCov_Setting open MeasureTheory Matrix
Formal statement
namespace MomentDRO.WorstCov
theorem exists_optimal_tight {n K : ℕ} [NeZero K] (a b : Fin K → ℝ) (x : Fin n → ℝ)
(μhat : Fin n → ℝ) (Sighat : Matrix (Fin n) (Fin n) ℝ) (hSig : Sighat.PosDef)
(γ2 : ℝ) (hγ2 : 0 < γ2) :
∃ (Lam : Fin K → Matrix (Fin n) (Fin n) ℝ) (lam : Fin K → Fin n → ℝ) (nu : Fin K → ℝ),
Feasible19 μhat Sighat γ2 Lam lam nu ∧
(∀ (Lam' : Fin K → Matrix (Fin n) (Fin n) ℝ) (lam' : Fin K → Fin n → ℝ)
(nu' : Fin K → ℝ), Feasible19 μhat Sighat γ2 Lam' lam' nu' →
obj19 a b x lam' nu' ≤ obj19 a b x lam nu) ∧
∑ k, Lam k = γ2 • Sighat + Matrix.vecMulVec μhat μhat := by sorry
end MomentDRO.WorstCov
Source
Delage & Ye, Distributionally robust optimization under moment uncertainty with application to data-driven problems, authors' draft of 20 Feb 2008 (OR 58(3), 2010), p. 20, §5.2, proof of Proposition 3, after (19d)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.