Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2.1 (Fenchel–Moreau), p. 435 — ρ** = lsc(ρ) for convex proper ρ (with lsc(ρ) > −∞)

Proved
ConvexRiskFn.Dual.theorem_2_1

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

biconjugateconvex-conjugatefenchel-moreaulower-semicontinuous-hullp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Let (X,Y)(\mathcal X,\mathcal Y)(X,Y) be a paired system as in §2 and ρ:X→R‾\rho:\mathcal X\to\overline{\mathbb R}ρ:X→R a convex (A1) and proper risk function whose lower semicontinuous hull lsc⁡(ρ)(X)=lim inf⁡Z→Xρ(Z)\operatorname{lsc}(\rho)(X)=\liminf_{Z\to X}\rho(Z)lsc(ρ)(X)=liminfZ→X​ρ(Z) is >−∞>-\infty>−∞ at every point. Then the biconjugate equals the lower semicontinuous hull:

ρ∗∗(X)=sup⁡μ∈Y{⟨μ,X⟩−ρ∗(μ)}=lsc⁡(ρ)(X)for all X∈X.\rho^{**}(X)=\sup_{\mu\in\mathcal Y}\{\langle\mu,X\rangle-\rho^*(\mu)\}=\operatorname{lsc}(\rho)(X)\qquad\text{for all }X\in\mathcal X.ρ∗∗(X)=μ∈Ysup​{⟨μ,X⟩−ρ∗(μ)}=lsc(ρ)(X)for all X∈X.

This is the Fenchel–Moreau theorem in the paired-space setting, cited in the paper (Rockafellar, Aubin–Ekeland) and used to derive the representation (2.5)–(2.6).

Formalization Note The hypothesis lsc⁡(ρ)>−∞\operatorname{lsc}(\rho)>-\inftylsc(ρ)>−∞ is added. Without it the statement as printed fails in infinite dimension: for a closed proper subspace LLL, a discontinuous linear ℓ\ellℓ on LLL, and ρ=ℓ\rho=\ellρ=ℓ on LLL, +∞+\infty+∞ off LLL, one has lsc⁡(ρ)=−∞\operatorname{lsc}(\rho)=-\inftylsc(ρ)=−∞ on LLL and +∞+\infty+∞ off LLL, while ρ∗≡+∞\rho^*\equiv+\inftyρ∗≡+∞ and ρ∗∗≡−∞\rho^{**}\equiv-\inftyρ∗∗≡−∞. The cited sources state the theorem with the closure cl⁡ρ\operatorname{cl}\rhoclρ, which coincides with lsc⁡(ρ)\operatorname{lsc}(\rho)lsc(ρ) exactly when lsc⁡(ρ)\operatorname{lsc}(\rho)lsc(ρ) never takes the value −∞-\infty−∞.

Preamble
import Mathlib
import Definitions.Def_ConvexRiskFn_Dual_Setting
Formal statement
namespace ConvexRiskFn.Dual

/-- Theorem 2.1 (Fenchel–Moreau), p. 435: for a convex proper `ρ : 𝒳 → ℝ̄` whose lower
semicontinuous hull never takes the value `−∞`, `ρ** = lsc(ρ)`. -/
theorem theorem_2_1 {Ω : Type*} [MeasurableSpace Ω] {𝒳 : Type*} [AddCommGroup 𝒳] [Module ℝ 𝒳]
    [TopologicalSpace 𝒳] [IsTopologicalAddGroup 𝒳] [ContinuousSMul ℝ 𝒳] [LocallyConvexSpace ℝ 𝒳]
    (S : PairedSpaces Ω 𝒳) (ρ : 𝒳 → EReal) (hcvx : A1 ρ) (hp : IsProper ρ)
    (hlscp : ∀ X, ⊥ < lscHull ρ X) :
    biconj S ρ = lscHull ρ := by sorry
end ConvexRiskFn.Dual
Source
Ruszczyński, Shapiro, Optimization of convex risk functions, Math. Oper. Res. 31 (2006), p. 435, Theorem 2.1 (Fenchel–Moreau), citing Rockafellar [22, Theorem 5] and Aubin–Ekeland [2, Theorem 4.4.2]
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 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