An L^p-Cauchy family indexed by ε → 0⁺ converges in L^p
ProvedDbarRegularity.exists_memLp_tendsto_eLpNorm_subfunctional-analysislp-spacesmeasure-theory
Let be a measure space, a Banach space and . Let be functions with for every , and suppose that for every there is with whenever . Then there is with
This is completeness of for families indexed by .
Formalization Note. is in ℝ≥0∞ with [Fact (1 ≤ p)]; the family is u : ℝ → α → E and its values at are ignored. The limit is along 𝓝[>] 0, stated as Tendsto (fun ε => eLpNorm (u ε - g) p μ) (𝓝[>] 0) (𝓝 0); the Cauchy condition is eLpNorm (u ε - u η) p μ ≤ ENNReal.ofReal δ.
Preamble
import Mathlib.MeasureTheory.Function.LpSpace.Complete import Mathlib.Topology.MetricSpace.Cauchy
Formal statement
namespace DbarRegularity
open MeasureTheory Filter Topology Set
open scoped ENNReal
/-- Completeness of `L^p` along `ε → 0⁺`: a family `u ε ∈ L^p` (`ε > 0`) which is Cauchy in
`L^p` as `ε → 0⁺` converges in `L^p` to some `g ∈ L^p`. -/
theorem exists_memLp_tendsto_eLpNorm_sub {α E : Type*} [MeasurableSpace α] {μ : Measure α}
[NormedAddCommGroup E] [CompleteSpace E] {p : ℝ≥0∞} [Fact (1 ≤ p)] (u : ℝ → α → E)
(hu : ∀ ε : ℝ, 0 < ε → MemLp (u ε) p μ)
(hc : ∀ δ : ℝ, 0 < δ → ∃ η₀ : ℝ, 0 < η₀ ∧ ∀ ε η : ℝ, 0 < ε → ε < η₀ → 0 < η → η < η₀ →
eLpNorm (u ε - u η) p μ ≤ ENNReal.ofReal δ) :
∃ g : α → E, MemLp g p μ ∧
Tendsto (fun ε : ℝ => eLpNorm (u ε - g) p μ) (𝓝[>] 0) (𝓝 0) := by sorry
end DbarRegularity
Source
E. M. Stein, Singular Integrals and Differentiability Properties of Functions, Princeton Mathematical Series 30, Princeton University Press (1970), Chapter II, § 3.2, p. 35