§5.12, Lemma 1 — Properties of the Minkowski gauge
ProvedVectorSpaceOpt.minkowski_gauge_propertiesLet be a real normed space and a convex set with ∈ interior . Then Mathlib's Minkowski gauge gauge K is nonnegative, positively homogeneous for positive real scalars, subadditive, and continuous. Its unit sublevel and strict sublevel recover the closure and interior of K:
This bundled lemma supplies the sublinear and topological data needed by the chapter's geometric separation results and is reusable whenever convex neighborhoods are represented by gauges.
import Mathlib
namespace VectorSpaceOpt
theorem minkowski_gauge_properties
{X : Type*} [NormedAddCommGroup X] [NormedSpace ℝ X]
(K : Set X) (hK : Convex ℝ K) (h₀ : (0 : X) ∈ interior K) :
(∀ x, 0 ≤ gauge K x) ∧
(∀ (a : ℝ), 0 < a → ∀ x, gauge K (a • x) = a * gauge K x) ∧
(∀ x₁ x₂, gauge K (x₁ + x₂) ≤ gauge K x₁ + gauge K x₂) ∧
Continuous (gauge K) ∧
{x | gauge K x ≤ 1} = closure K ∧
{x | gauge K x < 1} = interior K := by sorry
end VectorSpaceOptRead-back
What the Lean code literally says, in plain math · gpt-5
Let be any real normed vector space and let be convex with in its interior. For every , let be Mathlib's real Minkowski gauge, the infimum of the positive real numbers for which . The theorem asserts all six of the following simultaneously: for every ; for every real and every , ; for every , ; the total function is continuous; the set of points with gauge at most is exactly the closure of ; and the set of points with gauge strictly less than is exactly the interior of . Positive homogeneity is asserted only for , not for or negative ; is not assumed closed, bounded, or balanced, and is not assumed complete or nontrivial.
Confirmed by the mission captain (proposal self-audit).