§5.2, proof of Proposition 3, p. 20 — if [Λ λ; λᵀ ν] ⪰ 0 and ν > 0, a random vector with E[ζ] = λ/ν and E[ζζᵀ] = Λ/ν exists
ProvedMomentDRO.WorstCov.exists_law_of_bordered_psdmoment-problemp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1positive-semidefiniteschur-complement
Let , and be such that
Then there is a random vector (a Borel probability distribution on with finite second moments) with
In the proof of Proposition 3 this is applied to each block of an optimal solution of (19) to produce the vectors whose mixture is the worst-case distribution.
Formalization Note is the uncentred second moment, secondMomentAbout P 0.
Preamble
import Mathlib import Definitions.Def_MomentDRO_WorstCov_Setting open MeasureTheory Matrix
Formal statement
namespace MomentDRO.WorstCov
theorem exists_law_of_bordered_psd {n : ℕ} (L : Matrix (Fin n) (Fin n) ℝ) (l : Fin n → ℝ)
(v : ℝ) (hv : 0 < v) (hpsd : (bordered L l v).PosSemidef) :
∃ P : Measure (Fin n → ℝ), MomentDRO.Conf.HasSecondMoments P ∧
MomentDRO.Conf.meanVec P = v⁻¹ • l ∧ MomentDRO.Conf.secondMomentAbout P 0 = v⁻¹ • L := 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, construction of ζ_1, …, ζ_K
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.