§2, p. 211 — the restriction of to is
ProvedRockafellarMaxMono.Maximality.biconj_restrict_eqLet be a real Banach space, regarded as a subspace of its bidual through the canonical embedding , . If is a lower semicontinuous proper convex function on , then the biconjugate , a function on , satisfies
This is the Fenchel–Moreau theorem in the Banach-space setting, which the paper cites from Moreau. It is used, together with (2.2) and (2.4), to relate and .
Formalization Note The canonical embedding is Mathlib's NormedSpace.inclusionInDoubleDual ℝ E. is not assumed reflexive.
import Mathlib import Definitions.Def_RockafellarMaxMono_Shared_ProperConvex import Definitions.Def_RockafellarMaxMono_Shared_Conj
namespace RockafellarMaxMono.Maximality
theorem biconj_restrict_eq {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E]
[CompleteSpace E] (f : E → EReal) (hf : Shared.ProperConvex f) (hlsc : LowerSemicontinuous f) :
∀ x : E, Shared.conj (Shared.conj f) (NormedSpace.inclusionInDoubleDual ℝ E x) = f x := by sorry
end RockafellarMaxMono.Maximality
Read-back
What the Lean code literally says, in plain math · claude-opus-5-5
Setting. is a real Banach space, is its continuous dual, and is the continuous dual of . The canonical embedding is given by . The function has two hypotheses:
- satisfies the imported predicate
Shared.ProperConvex. It is not shown here; the name suggests "proper convex". - is lower semicontinuous for the norm topology.
The imported operation Shared.conj is not shown here; the name suggests the Fenchel conjugate. Applied to it gives . Applied again, to on the normed space , it gives .
Claim. For every ,
Degenerate cases. The claim concerns only points of in the image of . It says nothing about at other points of . If , the claim reduces to .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.