Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 9 — (P-Compactness) and (P-USC) imply a primal optimizer exists

Proved
ModelRiskOT.PrimalOpt.proposition_9

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

distributionally-robust-optimizationexistenceoptimal-transportp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1

Let SSS be a Polish space, μ\muμ a probability measure on SSS, δ>0\delta>0δ>0, and let ccc and fff satisfy (A1) and (A2). Assume the standing assumption of §5: (λ∗,φλ∗)∈Λc,f(\lambda^*,\varphi_{\lambda^*})\in\Lambda_{c,f}(λ∗,φλ∗​)∈Λc,f​ is a dual optimal pair with

I=J=J(λ∗,φλ∗)<∞.I=J=J(\lambda^*,\varphi_{\lambda^*})<\infty.I=J=J(λ∗,φλ∗​)<∞.

If (c,f)(c,f)(c,f) has Property (P-Compactness) at λ∗\lambda^*λ∗ and Property (P-USC), then there is a primal optimizer π∗∈Φμ,δ\pi^*\in\Phi_{\mu,\delta}π∗∈Φμ,δ​ with

I(π∗)=I=J=J(λ∗,φλ∗).I(\pi^*)=I=J=J(\lambda^*,\varphi_{\lambda^*}).I(π∗)=I=J=J(λ∗,φλ∗​).

The primal feasible set is in general not weakly compact, and the supremum need not be attained (Example 2 of the paper); the two properties replace compactness and upper semicontinuity.

Formalization Note The equality I=JI=JI=J is Theorem 1(a) of the paper, which is not assumed proved here; it enters as the hypothesis of the section's standing assumption, together with the dual optimality J=J(λ∗,φλ∗)J=J(\lambda^*,\varphi_{\lambda^*})J=J(λ∗,φλ∗​) and finiteness. No sign condition on λ∗\lambda^*λ∗ beyond λ∗≥0\lambda^*\ge0λ∗≥0 (part of Λc,f\Lambda_{c,f}Λc,f​) is assumed.

Preamble
import Mathlib
import Definitions.Def_ModelRiskOT_PrimalOpt_PCompactness
import Definitions.Def_ModelRiskOT_PrimalOpt_PUSC
Formal statement
namespace ModelRiskOT.PrimalOpt

open MeasureTheory

/-- **Proposition 9** (§5, p. 26). Assume (A1), (A2) and the standing assumption of §5:
`(λ*, φ_{λ*}) ∈ Λ_{c,f}` is a dual optimal pair with `I = J = J(λ*, φ_{λ*}) < ∞` (the identity
`I = J` is Theorem 1(a), taken here as a hypothesis). If (P-Compactness) holds for `λ*` and (P-USC)
holds, there is a primal optimizer `π* ∈ Φ_{μ,δ}` with `I(π*) = I = J = J(λ*, φ_{λ*})`. -/
theorem proposition_9 {S : Type*} [TopologicalSpace S] [PolishSpace S] [MeasurableSpace S]
    [BorelSpace S] (c : S → S → ℝ) (f : S → ℝ) (μ : Measure S) [IsProbabilityMeasure μ]
    (δ : ℝ) (hδ : 0 < δ) (hA1 : ModelRiskOT.Duality.AssumptionA1 c) (hA2 : AssumptionA2 f μ)
    (lam : ℝ) (hfeas : (lam, phiLam c f lam) ∈ dualFeasible c f)
    (hIJ : ModelRiskOT.Duality.primalValue c f μ δ = ModelRiskOT.Duality.dualValue c f μ δ)
    (hopt : ModelRiskOT.Duality.dualValue c f μ δ = ModelRiskOT.Duality.dualObj μ δ lam (phiLam c f lam))
    (hfin : ModelRiskOT.Duality.dualObj μ δ lam (phiLam c f lam) < ⊤)
    (hP1 : PCompactness c f μ lam) (hP2 : PUSC c f μ δ) :
    ∃ πstar ∈ ModelRiskOT.Duality.primalFeasible c μ δ,
      ModelRiskOT.Duality.primalObj f πstar = ModelRiskOT.Duality.primalValue c f μ δ ∧ ModelRiskOT.Duality.primalValue c f μ δ = ModelRiskOT.Duality.dualValue c f μ δ ∧
        ModelRiskOT.Duality.dualValue c f μ δ = ModelRiskOT.Duality.dualObj μ δ lam (phiLam c f lam) := by sorry

end ModelRiskOT.PrimalOpt
Source
Blanchet & Murthy, Quantifying Distributional Model Risk via Optimal Transport, arXiv:1604.01446v2, p. 26, Proposition 9 (with the standing assumption of §5)
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