Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 9, p. 39 — for Z ∈ [0, M], inf over η ∈ ℝ of g(η; P) equals the inf over η ∈ [−M/(c_k−1), M]

Proved
UniformDRO.Concentration.lemma_9

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

cressie-readdistributionally-robust-optimizationdualityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Let PPP be a probability measure on (X,A)(\mathcal X,\mathcal A)(X,A), k>1k>1k>1, ρ>0\rho>0ρ>0, k∗=k/(k−1)k_*=k/(k-1)k∗​=k/(k−1), ck=(1+k(k−1)ρ)1/kc_k=(1+k(k-1)\rho)^{1/k}ck​=(1+k(k−1)ρ)1/k, and let Z:X→[0,M]Z:\mathcal X\to[0,M]Z:X→[0,M] be measurable. With g(η;P)=ck EP[(Z−η)+k∗]1/k∗+ηg(\eta;P)=c_k\,\mathbb E_P[(Z-\eta)_+^{k_*}]^{1/k_*}+\etag(η;P)=ck​EP​[(Z−η)+k∗​​]1/k∗​+η,

inf⁡η∈Rg(η;P)=inf⁡{g(η;P):η∈[−1ck−1M, M]}.\inf_{\eta\in\mathbb R}g(\eta;P)=\inf\Big\{g(\eta;P):\eta\in\Big[-\frac{1}{c_k-1}M,\,M\Big]\Big\}.η∈Rinf​g(η;P)=inf{g(η;P):η∈[−ck​−11​M,M]}.

Restricting the dual variable to a compact interval is what makes a finite grid of η\etaη values, and hence a union bound, sufficient in the proof of Theorem 2.

Formalization Note Since k>1k>1k>1 and ρ>0\rho>0ρ>0 give ck>1c_k>1ck​>1, the interval is nonempty (−M/(ck−1)≤0≤M-M/(c_k-1)\le0\le M−M/(ck​−1)≤0≤M), and g(⋅;P)≥EP[Z]≥0g(\cdot;P)\ge\mathbb E_P[Z]\ge0g(⋅;P)≥EP​[Z]≥0, so both infima are real infima of sets bounded below. Measurability of ZZZ is added. No condition M≥1M\ge1M≥1 is imposed: Lemma 9 holds for any M≥0M\ge0M≥0 (and Z∈[0,M]Z\in[0,M]Z∈[0,M] forces M≥0M\ge0M≥0).

Preamble
import Mathlib
import Definitions.Def_UniformDRO_Concentration_RobustRisk

open MeasureTheory
Formal statement
namespace UniformDRO.Concentration

/-- Lemma 9 (Duchi & Namkoong, arXiv:1810.08750v6, App. C.1, p. 39): if `Z ∈ [0, M]`, then for any
probability measure `P`,
`inf_{η ∈ ℝ} g_k(η; P) = inf { g_k(η; P) : η ∈ [-M/(c_k - 1), M] }`.
Both infima are of functions bounded below (`g_k(η; P) ≥ E_P[Z] ≥ 0`), and the interval is
nonempty because `c_k > 1`, so `-M/(c_k - 1) ≤ 0 ≤ M`. -/
theorem lemma_9 {X : Type*} [MeasurableSpace X] (P : Measure X) [IsProbabilityMeasure P]
    (k ρ : ℝ) (hk : 1 < k) (hρ : 0 < ρ) (Z : X → ℝ) (hZm : Measurable Z) (M : ℝ)
    (hZ : ∀ x, Z x ∈ Set.Icc 0 M) :
    ⨅ η : ℝ, dualObjective k ρ P Z η =
      ⨅ η : Set.Icc (-(M / (ck k ρ - 1))) M, dualObjective k ρ P Z η := by sorry

end UniformDRO.Concentration
Source
Duchi & Namkoong, arXiv:1810.08750v6, Lemma 9, App. C.1, p. 39; proof pp. 39–40
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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