VectorSpaceOpt.restrictedConvexConjugate
DefinitionVectorSpaceOpt_restrictedConvexConjugateLet be a real normed space, , , and a continuous linear functional. The restricted convex conjugate is the real supremum
The definition itself uses sSup; theorem statements separately require to be nonempty and the displayed set to be bounded above whenever the value is used as a finite conjugate. This separation matches Luenberger's domain and prevents the default behavior of an unbounded real supremum from becoming an implicit mathematical assertion. The definition is reusable for closed-epigraph and duality statements on independent convex domains.
import Mathlib
namespace VectorSpaceOpt
/-- The convex conjugate of a finite-valued function restricted to a set. -/
noncomputable def restrictedConvexConjugate
{X : Type*} [NormedAddCommGroup X] [NormedSpace ℝ X]
(C : Set X) (f : X → ℝ) (ℓ : X →L[ℝ] ℝ) : ℝ :=
sSup ((fun x : X ↦ ℓ x - f x) '' C)
end VectorSpaceOptRead-back
What the Lean code literally says, in plain math · gpt-5
For every real normed vector space , set , real-valued function on , and continuous real-linear functional on , this defines the restricted convex conjugate to be the real supremum . This uses the total real-valued , not an extended-real supremum: when the displayed image is nonempty and bounded above it is the usual supremum, while if is empty or the image is unbounded above Mathlib's totalized real has value . No nonemptiness or boundedness hypothesis is built into the definition, and remains finite-valued.
Confirmed by the mission captain (proposal self-audit).