§5.2, proof of Proposition 3, pp. 20–21 — E[max_l −a_l xᵀξ* − b_l] ≥ Σ_k (−a_k xᵀλ_k − b_k ν_k)
ProvedMomentDRO.WorstCov.mixture_value_gedistributionally-robust-optimizationmixture-distributionp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1piecewise-linear-utility
Let , , , , , let be feasible for (19) with every , and let be distributions with finite second moments, and . Then the mixture satisfies
Combined with the bound of (18) by (19), this shows that the mixture built from an optimal solution of (19) attains the maximum of (18).
Formalization Note The objective is written with the sign of the page's displayed computation (p. 21), which is also the sign used throughout this mission.
Preamble
import Mathlib import Definitions.Def_MomentDRO_WorstCov_Setting open MeasureTheory Matrix
Formal statement
namespace MomentDRO.WorstCov
theorem mixture_value_ge {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)
(hfeas : Feasible19 μhat Sighat γ2 Lam lam nu)
(hpos : ∀ k, 0 < nu k)
(Pk : Fin K → Measure (Fin n → ℝ)) (hPk : ∀ k, MomentDRO.Conf.HasSecondMoments (Pk k))
(hmean : ∀ k, MomentDRO.Conf.meanVec (Pk k) = (nu k)⁻¹ • lam k)
(hsecond : ∀ k, MomentDRO.Conf.secondMomentAbout (Pk k) 0 = (nu k)⁻¹ • Lam k) :
obj19 a b x lam nu ≤ ∫ ξ, pwCost a b x ξ ∂(mixture nu Pk) := 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), pp. 20–21, §5.2, proof of Proposition 3, last display
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.