Interior of the Effective Domain from Finite Integral
ProvedFunctionalInequalities.Probability.mem_interior_setOf_ne_top_of_lintegral_ne_top_readableThis is the effective-domain interior lemma corresponding to Lemma 16 of Cordero–Erausquin–Klartag (arXiv:1304.0630v1).
Let be finite-dimensional, let be a finite Borel measure, and let lie in the interior of the convex hull of . Let satisfy:
- ;
- its real epigraph is convex;
- is almost everywhere measurable;
- .
Then
Thus finite extended-real integral control forces finiteness on a neighborhood of every interior support point.
Original Lean declaration header
Within namespace FunctionalInequalities.Probability. Shown in its original module context. Proof omitted.
theorem mem_interior_setOf_ne_top_of_lintegral_ne_top [FiniteDimensional ℝ E] [BorelSpace E] (μ : Measure E) [IsFiniteMeasure μ] (x₀ : E) :
x₀ ∈ interior (convexHull ℝ μ.support) →
∀ φ : E → EReal, (∀ x, 0 ≤ φ x) → Convex ℝ {p : E × ℝ | φ p.1 ≤ (p.2 : EReal)} →
AEMeasurable (fun x => (φ x).toENNReal) μ →
(∫⁻ x, (φ x).toENNReal ∂μ) ≠ ⊤ →
x₀ ∈ interior {x | φ x ≠ ⊤}
Module navigation
Module: Probability/ConvexSupport.
Previous: exists_pos_le_measure_halfspace_of_mem_interior_convexHull_support · Next: exists_pos_le_measure_halfspace_of_mem_interior_convexHull_support_of_isFiniteMeasure.
Source excerpt, lines 434–439 · Frozen source 408915f.
import Definitions.Def_KLSModular_FunctionalInequalities_Probability_ConvexSupport
open MeasureTheory Set
variable {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E]
[MeasurableSpace E]
theorem FunctionalInequalities.Probability.mem_interior_setOf_ne_top_of_lintegral_ne_top_readable [FiniteDimensional ℝ E] [BorelSpace E] (μ : Measure E) [IsFiniteMeasure μ] (x₀ : E) :
x₀ ∈ interior (convexHull ℝ μ.support) →
∀ φ : E → EReal, (∀ x, 0 ≤ φ x) → Convex ℝ {p : E × ℝ | φ p.1 ≤ (p.2 : EReal)} →
AEMeasurable (fun x => (φ x).toENNReal) μ →
(∫⁻ x, (φ x).toENNReal ∂μ) ≠ ⊤ →
x₀ ∈ interior {x | φ x ≠ ⊤} := by sorry