Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 1, p. 16 — Γ* ≠ ∅ and ℓ_rob < ∞ above κ√δβᵀA(x)⁻¹β, ℓ_rob = ∞ below; 𝕌₁ ⊆ 𝕌 ⊆ 𝕌₂

Proved
OTDRO.WorstCase.effective_domain

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

distributionally-robust-optimizationdualityeffective-domainp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Assume Assumptions 1 and 2, let δ>0\delta>0δ>0 and let B⊆RdB\subseteq\mathbb R^dB⊆Rd be convex. Write κ\kappaκ for the growth exponent of ℓ\ellℓ, Γ∗(β,λ;x)\Gamma^*(\beta,\lambda;x)Γ∗(β,λ;x) for the set of maximizers of γ↦F(γ,β,λ;x)\gamma\mapsto F(\gamma,\beta,\lambda;x)γ↦F(γ,β,λ;x), ℓrob=sup⁡γF\ell_{\rm rob}=\sup_\gamma Fℓrob​=supγ​F and fδ(β,λ)=EP0[ℓrob(β,λ;X)]f_\delta(\beta,\lambda)=E_{P_0}[\ell_{\rm rob}(\beta,\lambda;X)]fδ​(β,λ)=EP0​​[ℓrob​(β,λ;X)]. Then for every β∈B\beta\in Bβ∈B, λ≥0\lambda\ge0λ≥0 and x∈Rdx\in\mathbb R^dx∈Rd:

  1. if λ>κδ βTA(x)−1β\lambda>\kappa\sqrt\delta\,\beta^{\mathsf T}A(x)^{-1}\betaλ>κδ​βTA(x)−1β, then Γ∗(β,λ;x)≠∅\Gamma^*(\beta,\lambda;x)\neq\emptysetΓ∗(β,λ;x)=∅ and ℓrob(β,λ;x)<∞\ell_{\rm rob}(\beta,\lambda;x)<\inftyℓrob​(β,λ;x)<∞;
  2. if λ<κδ βTA(x)−1β\lambda<\kappa\sqrt\delta\,\beta^{\mathsf T}A(x)^{-1}\betaλ<κδ​βTA(x)−1β, then Γ∗(β,λ;x)=∅\Gamma^*(\beta,\lambda;x)=\emptysetΓ∗(β,λ;x)=∅ and ℓrob(β,λ;x)=∞\ell_{\rm rob}(\beta,\lambda;x)=\inftyℓrob​(β,λ;x)=∞.

Consequently, with U={(β,λ)∈B×R+:fδ(β,λ)<∞}\mathbb U=\{(\beta,\lambda)\in B\times\mathbb R_+: f_\delta(\beta,\lambda)<\infty\}U={(β,λ)∈B×R+​:fδ​(β,λ)<∞},

U1={λ>λthr(β)} ⊆ U ⊆ U2={λ≥λthr(β)}.\mathbb U_1=\{\lambda>\lambda_{thr}(\beta)\}\ \subseteq\ \mathbb U\ \subseteq\ \mathbb U_2=\{\lambda\ge\lambda_{thr}(\beta)\}.U1​={λ>λthr​(β)} ⊆ U ⊆ U2​={λ≥λthr​(β)}.

The lemma identifies the effective domain of the dual objective up to its boundary λ=λthr(β)\lambda=\lambda_{thr}(\beta)λ=λthr​(β); it is the reason a dual optimizer always satisfies λ∗(β)≥λthr(β)\lambda_*(\beta)\ge\lambda_{thr}(\beta)λ∗​(β)≥λthr​(β) and the region where the worst case of Theorem 6(c) is analysed.

Formalization Note The page prints the threshold as "κδβA(x)−1β\kappa\sqrt\delta\beta A(x)^{-1}\betaκδ​βA(x)−1β", a typo for βTA(x)−1β\beta^{\mathsf T}A(x)^{-1}\betaβTA(x)−1β. The inclusions are stated pointwise: for β∈B\beta\in Bβ∈B and λ≥0\lambda\ge0λ≥0, λ>λthr(β)\lambda>\lambda_{thr}(\beta)λ>λthr​(β) implies fδ(β,λ)<∞f_\delta(\beta,\lambda)<\inftyfδ​(β,λ)<∞, and fδ(β,λ)<∞f_\delta(\beta,\lambda)<\inftyfδ​(β,λ)<∞ implies λ≥λthr(β)\lambda\ge\lambda_{thr}(\beta)λ≥λthr​(β). ℓrob\ell_{\rm rob}ℓrob​ is a supremum of real numbers over R\mathbb RR, so it is never −∞-\infty−∞, and "finite" is written ℓrob≠+∞\ell_{\rm rob}\ne+\inftyℓrob​=+∞. The same statement is drafted in the sibling mission on convexity of the dual objective.

Preamble
import Mathlib
import Definitions.Def_OTDRO_Dual_Setting

open MeasureTheory
Formal statement
namespace OTDRO.WorstCase

/-- **Lemma 1** (arXiv:1810.02403v3, p. 16; proof p. 47). (a) If `λ > κ√δ βᵀA(x)⁻¹β` then
`Γ*(β, λ; x)` is nonempty and `ℓ_rob(β, λ; x)` is finite; (b) if `λ < κ√δ βᵀA(x)⁻¹β` then
`Γ*(β, λ; x) = ∅` and `ℓ_rob(β, λ; x) = ∞`; consequently `𝕌₁ ⊆ 𝕌 ⊆ 𝕌₂`, where
`𝕌 = {(β, λ) ∈ B × ℝ₊ : f_δ(β, λ) < ∞}`, `𝕌₁` (resp. `𝕌₂`) is `λ > λ_thr(β)` (resp. `≥`). -/
theorem effective_domain {d : ℕ} (P0 : Measure (EuclideanSpace ℝ (Fin d)))
    [IsProbabilityMeasure P0] (A : EuclideanSpace ℝ (Fin d) → Matrix (Fin d) (Fin d) ℝ)
    (ρmin ρmax : ℝ) (hA : OTDRO.Dual.Assumption1 P0 A ρmin ρmax) (ℓ : ℝ → ℝ) (h2 : OTDRO.Dual.Assumption2 P0 ℓ)
    (δ : ℝ) (hδ : 0 < δ) (B : Set (EuclideanSpace ℝ (Fin d))) (hB : Convex ℝ B) :
    (∀ β ∈ B, ∀ lam : ℝ, 0 ≤ lam → ∀ x : EuclideanSpace ℝ (Fin d),
      (OTDRO.Dual.growthKappa ℓ * Real.sqrt δ * OTDRO.Dual.quadInv A β x < lam →
        (OTDRO.Dual.maximizers ℓ A δ β lam x).Nonempty ∧ OTDRO.Dual.ellRob ℓ A δ β lam x ≠ ⊤) ∧
      (lam < OTDRO.Dual.growthKappa ℓ * Real.sqrt δ * OTDRO.Dual.quadInv A β x →
        OTDRO.Dual.maximizers ℓ A δ β lam x = ∅ ∧ OTDRO.Dual.ellRob ℓ A δ β lam x = ⊤)) ∧
    (∀ β ∈ B, ∀ lam : ℝ, 0 ≤ lam → OTDRO.Dual.lamThr P0 ℓ A δ β < lam → OTDRO.Dual.fDelta P0 ℓ A δ β lam < ⊤) ∧
    (∀ β ∈ B, ∀ lam : ℝ, 0 ≤ lam → OTDRO.Dual.fDelta P0 ℓ A δ β lam < ⊤ → OTDRO.Dual.lamThr P0 ℓ A δ β ≤ lam) := by sorry

end OTDRO.WorstCase
Source
arXiv:1810.02403v3, Lemma 1, p. 16; (11), p. 15; proof p. 47
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me