Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proof of Theorem 4.2, (12d)–(12e), p. 13 — under Assumption 4.1, the minimax interchange in each constraint is exact

Proved
WassDDRO.Reduction.minimax_12e

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

convex-analysisdual-normminimaxp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1wasserstein

Let EEE be a finite-dimensional real normed space with dual norm ∥⋅∥∗\|\cdot\|_*∥⋅∥∗​, and let Ξ\XiΞ, ℓ1,…,ℓK\ell_1,\dots,\ell_Kℓ1​,…,ℓK​ and samples ξ^1,…,ξ^N∈Ξ\hat\xi_1,\dots,\hat\xi_N\in\Xiξ^​1​,…,ξ^​N​∈Ξ be as in Theorem 4.2, with Assumption 4.1 in force. Fix λ≥0\lambda\ge0λ≥0, a sample index iii and a piece kkk. Then

sup⁡ξ∈Ξ(ℓk(ξ)−λ∥ξ−ξ^i∥)=min⁡∥z∥∗≤λ sup⁡ξ∈Ξ(ℓk(ξ)−⟨z,ξ−ξ^i⟩),\sup_{\xi\in\Xi}\big(\ell_k(\xi)-\lambda\|\xi-\hat\xi_i\|\big)=\min_{\|z\|_*\le\lambda}\ \sup_{\xi\in\Xi}\big(\ell_k(\xi)-\langle z,\xi-\hat\xi_i\rangle\big),ξ∈Ξsup​(ℓk​(ξ)−λ∥ξ−ξ^​i​∥)=∥z∥∗​≤λmin​ ξ∈Ξsup​(ℓk​(ξ)−⟨z,ξ−ξ^​i​⟩),

and the minimum on the right is attained by some zzz with ∥z∥∗≤λ\|z\|_*\le\lambda∥z∥∗​≤λ.

Since λ∥ξ−ξ^i∥=max⁡∥z∥∗≤λ⟨z,ξ−ξ^i⟩\lambda\|\xi-\hat\xi_i\|=\max_{\|z\|_*\le\lambda}\langle z,\xi-\hat\xi_i\rangleλ∥ξ−ξ^​i​∥=max∥z∥∗​≤λ​⟨z,ξ−ξ^​i​⟩, the left side is the constraint function of (12d) and the right side that of (12e); the statement says that the inequality between (12d) and (12e) is an equality, by the minimax theorem over the compact dual ball.

Formalization Note Both sides are extended reals; the infimum over the dual ball is written as an infimum over the set {z:∥z∥∗≤λ}\{z:\|z\|_*\le\lambda\}{z:∥z∥∗​≤λ} and its attainment is stated separately. The left side is written with λ∥ξ−ξ^i∥\lambda\|\xi-\hat\xi_i\|λ∥ξ−ξ^​i​∥ rather than with the inner maximum over zzz.

Preamble
import Mathlib
import Definitions.Def_WassDDRO_Reduction_Setting

open MeasureTheory Filter Topology
Formal statement
namespace WassDDRO.Reduction

/-- Proof of Theorem 4.2, (12d)–(12e), p. 13: under Assumption 4.1, for each `λ ≥ 0`, `i`, `k`,
the constraint function of (12d) equals that of (12e) (minimax), and the minimum over the dual
ball `‖z‖_* ≤ λ` is attained. -/
theorem minimax_12e {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E]
    [MeasurableSpace E] [BorelSpace E]
    {K N : ℕ} (hK : 0 < K) (hN : 0 < N) (Ξ : Set E) (ℓ : Fin K → E → EReal)
    (hmeas : ∀ k, Measurable (ℓ k)) (ξhat : Fin N → E) (hξ : ∀ i, ξhat i ∈ Ξ)
    (hA : Assumption41 Ξ ℓ) (lam : ℝ) (hlam : 0 ≤ lam) (i : Fin N) (k : Fin K) :
    (⨆ ξ ∈ Ξ, ℓ k ξ - ((lam * ‖ξ - ξhat i‖ : ℝ) : EReal))
        = (⨅ z ∈ {z : StrongDual ℝ E | ‖z‖ ≤ lam},
            ⨆ ξ ∈ Ξ, ℓ k ξ - ((z (ξ - ξhat i) : ℝ) : EReal)) ∧
      ∃ z : StrongDual ℝ E, ‖z‖ ≤ lam ∧
        (⨆ ξ ∈ Ξ, ℓ k ξ - ((z (ξ - ξhat i) : ℝ) : EReal))
          = (⨆ ξ ∈ Ξ, ℓ k ξ - ((lam * ‖ξ - ξhat i‖ : ℝ) : EReal)) := by sorry

end WassDDRO.Reduction
Source
Mohajerin Esfahani & Kuhn, arXiv:1505.05116v3, proof of Theorem 4.2, (12d) p. 12 and (12e) p. 13; minimax step p. 13
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