Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3.6, p. 8 — asymptotic consistency: P^∞-a.s. Ĵ_N → J⋆, and accumulation points of x̂_N are optimal for (1)

Proved
WassDDRO.Consistency.theorem_3_6

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

asymptotic-consistencydistributionally-robust-optimizationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1wasserstein

Let EEE be a finite-dimensional real normed space with its Borel σ\sigmaσ-algebra, m=dim⁡E≠2m=\dim E\ne2m=dimE=2, and PPP a probability distribution on EEE supported on Ξ⊆E\Xi\subseteq EΞ⊆E. Suppose Assumption 3.3 holds: for some a>1a>1a>1, A=EP[exp⁡(∥ξ∥a)]<∞A=\mathbb E^P[\exp(\|\xi\|^a)]<\inftyA=EP[exp(∥ξ∥a)]<∞; let c1,c2>0c_1,c_2>0c1​,c2​>0 be constants for which the concentration inequality (7) holds, and εN(β)\varepsilon_N(\beta)εN​(β) the radius (8). Let βN∈(0,1)\beta_N\in(0,1)βN​∈(0,1), N∈NN\in\mathbb NN∈N, satisfy

∑N=1∞βN<∞andlim⁡N→∞εN(βN)=0.\sum_{N=1}^\infty\beta_N<\infty\qquad\text{and}\qquad\lim_{N\to\infty}\varepsilon_N(\beta_N)=0 .N=1∑∞​βN​<∞andN→∞lim​εN​(βN​)=0.

Let X⊆RnX\subseteq\mathbb R^nX⊆Rn, and h:Rn×E→Rh:\mathbb R^n\times E\to\mathbb Rh:Rn×E→R with h(x,⋅)h(x,\cdot)h(x,⋅) measurable for every xxx. Let ξ^1,ξ^2,…\hat\xi_1,\hat\xi_2,\ldotsξ^​1​,ξ^​2​,… be i.i.d. with law PPP (joint law P∞P^\inftyP∞), and for each N≥1N\ge1N≥1 let x^N∈X\widehat x_N\in XxN​∈X be an optimizer, and J^N\widehat J_NJN​ the optimal value, of the distributionally robust program (5) with ambiguity set BεN(βN)(P^N)\mathbb B_{\varepsilon_N(\beta_N)}(\widehat P_N)BεN​(βN​)​(PN​) built from the first NNN samples. Let J⋆=inf⁡x∈XEP[h(x,ξ)]J^\star=\inf_{x\in X}\mathbb E^P[h(x,\xi)]J⋆=infx∈X​EP[h(x,ξ)] be the optimal value of (1).

  1. If h(x,ξ)h(x,\xi)h(x,ξ) is upper semicontinuous in ξ∈Ξ\xi\in\Xiξ∈Ξ for every x∈Xx\in Xx∈X, and there is L≥0L\ge0L≥0 with
∣h(x,ξ)∣≤L(1+∥ξ∥)for all x∈X, ξ∈Ξ,|h(x,\xi)|\le L(1+\|\xi\|)\qquad\text{for all }x\in X,\ \xi\in\Xi,∣h(x,ξ)∣≤L(1+∥ξ∥)for all x∈X, ξ∈Ξ,

then P∞P^\inftyP∞-almost surely J^N≥J⋆\widehat J_N\ge J^\starJN​≥J⋆ for all sufficiently large NNN and J^N→J⋆\widehat J_N\to J^\starJN​→J⋆ as N→∞N\to\inftyN→∞. 2. If the assumptions of (1) hold, XXX is closed, and h(x,ξ)h(x,\xi)h(x,ξ) is lower semicontinuous in x∈Xx\in Xx∈X for every ξ∈Ξ\xi\in\Xiξ∈Ξ, then P∞P^\inftyP∞-almost surely every accumulation point x⋆x^\starx⋆ of (x^N)N∈N(\widehat x_N)_{N\in\mathbb N}(xN​)N∈N​ is an optimal solution of (1): x⋆∈Xx^\star\in Xx⋆∈X and EP[h(x⋆,ξ)]≤EP[h(x,ξ)]\mathbb E^P[h(x^\star,\xi)]\le\mathbb E^P[h(x,\xi)]EP[h(x⋆,ξ)]≤EP[h(x,ξ)] for all x∈Xx\in Xx∈X.

With radii shrinking at a rate dictated by the summable confidence levels βN\beta_NβN​ (for example βN=exp⁡(−N)\beta_N=\exp(-\sqrt N)βN​=exp(−N​)), the Wasserstein distributionally robust certificates and decisions are therefore consistent for the true stochastic program.

Formalization Note The paper writes "J^N↓J⋆\widehat J_N\downarrow J^\starJN​↓J⋆"; its proof establishes J⋆≤J^NJ^\star\le\widehat J_NJ⋆≤JN​ for all large NNN and lim sup⁡J^N≤J⋆\limsup\widehat J_N\le J^\starlimsupJN​≤J⋆, which is what is stated (convergence from above, not monotonicity). The constants c1,c2c_1,c_2c1​,c2​ are any constants satisfying (7), which Theorem 3.4 supplies under Assumption 3.3 for m≠2m\ne2m=2; (7) and (8) are printed only for m≠2m\ne2m=2. The loss is real-valued (the paper allows extended values, but the growth bound makes hhh finite on X×ΞX\times\XiX×Ξ). The optimizers x^N\widehat x_NxN​ are required to exist only for samples in Ξ\XiΞ, which holds P∞P^\inftyP∞-almost surely. The worst-case expectation counts distributions under which h(x,⋅)h(x,\cdot)h(x,⋅) is integrable; under the growth bound every distribution in the ball integrates h(x,⋅)h(x,\cdot)h(x,⋅), and Assumption 3.3 makes the Bochner integrals EP[h(x,ξ)]\mathbb E^P[h(x,\xi)]EP[h(x,ξ)] genuine. "Accumulation point" is a cluster point of the sequence along N→∞N\to\inftyN→∞; values of x^0\widehat x_0x0​ are irrelevant.

Preamble
import Mathlib
import Definitions.Def_WassDDRO_Consistency_Setting

open MeasureTheory Filter Topology
Formal statement
namespace WassDDRO.Consistency

/-- Theorem 3.6 (Asymptotic consistency), p. 8. Under Assumption 3.3, with constants `c₁, c₂`
for which (7) holds (`m ≠ 2`), `β_N ∈ (0, 1)` summable with `ε_N(β_N) → 0`, and `x̂_N` an
optimizer of (5) with ambiguity set `B_{ε_N(β_N)}(P̂_N)`:
(i) if `h(x, ·)` is upper semicontinuous on `Ξ` and `|h(x, ξ)| ≤ L(1 + ‖ξ‖)` on `X × Ξ`, then
`P^∞`-a.s. `Ĵ_N ≥ J⋆` eventually and `Ĵ_N → J⋆`;
(ii) if moreover `X` is closed and `h(·, ξ)` is lower semicontinuous on `X` for every `ξ ∈ Ξ`,
then `P^∞`-a.s. every accumulation point of `(x̂_N)` is an optimal solution of (1). -/
theorem theorem_3_6 {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E]
    [MeasurableSpace E] [BorelSpace E]
    (P : Measure E) [IsProbabilityMeasure P] (Ξ : Set E) (hP : P Ξᶜ = 0)
    (a : ℝ) (ha : 1 < a) (hA : Integrable (fun ξ => Real.exp (‖ξ‖ ^ a)) P)
    (hm : Module.finrank ℝ E ≠ 2)
    (c₁ c₂ : ℝ) (hc₁ : 0 < c₁) (hc₂ : 0 < c₂) (h7 : Concentration7 P a c₁ c₂)
    (β : ℕ → ℝ) (hβ : ∀ N, 0 < β N ∧ β N < 1) (hsum : Summable β)
    (hlim : Tendsto (fun N => radius c₁ c₂ a (Module.finrank ℝ E) N (β N)) atTop (𝓝 0))
    {n : ℕ} (X : Set (EuclideanSpace ℝ (Fin n))) (h : EuclideanSpace ℝ (Fin n) → E → ℝ)
    (hmeas : ∀ x, Measurable (h x))
    (xhat : (N : ℕ) → (Fin N → E) → EuclideanSpace ℝ (Fin n))
    (hopt : ∀ (N : ℕ) (ξhat : Fin N → E), 1 ≤ N → (∀ i, ξhat i ∈ Ξ) →
      xhat N ξhat ∈ X ∧ ∀ x ∈ X,
        WassersteinDRO.Duality.worstCaseRisk (radius c₁ c₂ a (Module.finrank ℝ E) N (β N)) 1 Ξ
            (WassersteinDRO.Duality.empiricalDistribution ξhat) (h (xhat N ξhat)) ≤
          WassersteinDRO.Duality.worstCaseRisk (radius c₁ c₂ a (Module.finrank ℝ E) N (β N)) 1 Ξ
            (WassersteinDRO.Duality.empiricalDistribution ξhat) (h x)) :
    ((∀ x ∈ X, UpperSemicontinuousOn (h x) Ξ) → ∀ L : ℝ, 0 ≤ L →
        (∀ x ∈ X, ∀ ξ ∈ Ξ, |h x ξ| ≤ L * (1 + ‖ξ‖)) →
        ∀ᵐ ω ∂(P_inf P),
          (∀ᶠ N in atTop, trueValue X h P ≤
            droValue X h (radius c₁ c₂ a (Module.finrank ℝ E) N (β N)) Ξ (fun i : Fin N => ω i)) ∧
          Tendsto (fun N => droValue X h (radius c₁ c₂ a (Module.finrank ℝ E) N (β N)) Ξ
            (fun i : Fin N => ω i)) atTop (𝓝 (trueValue X h P))) ∧
    ((∀ x ∈ X, UpperSemicontinuousOn (h x) Ξ) → ∀ L : ℝ, 0 ≤ L →
        (∀ x ∈ X, ∀ ξ ∈ Ξ, |h x ξ| ≤ L * (1 + ‖ξ‖)) →
        IsClosed X → (∀ ξ ∈ Ξ, LowerSemicontinuousOn (fun x => h x ξ) X) →
        ∀ᵐ ω ∂(P_inf P), ∀ xstar : EuclideanSpace ℝ (Fin n),
          MapClusterPt xstar atTop (fun N => xhat N (fun i : Fin N => ω i)) →
          xstar ∈ X ∧ ∀ x ∈ X, ∫ ξ, h xstar ξ ∂P ≤ ∫ ξ, h x ξ ∂P) := by sorry

end WassDDRO.Consistency
Source
Mohajerin Esfahani & Kuhn, arXiv:1505.05116v3, Theorem 3.6, p. 8; proof pp. 9–10
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