Young's inequality on (squared lower-integral form)
ProvedDbarRegularity.lintegral_sq_lintegral_mul_leconvolutionharmonic-analysislp-spacesmeasure-theory
Let and let be almost-everywhere strongly measurable. Then
In other words , the case of Young's convolution inequality, stated for absolute values and without any finiteness assumptions.
Besides the norm bound for convolutions, it shows that when and the integral is finite for almost every .
Formalization Note. is EuclideanSpace ℝ (Fin d) with Lebesgue measure volume. All integrals are lower Lebesgue integrals ∫⁻ of extended nonnegative reals, absolute values are ‖·‖ₑ, and squares are natural-number powers in ℝ≥0∞. Only AEStronglyMeasurable is assumed; both sides may be infinite.
Preamble
import Mathlib.MeasureTheory.Integral.MeanInequalities import Mathlib.MeasureTheory.Function.LpSeminorm.Basic import Mathlib.MeasureTheory.Group.Prod import Mathlib.MeasureTheory.Measure.Haar.InnerProductSpace import Mathlib.Analysis.InnerProductSpace.PiL2
Formal statement
namespace DbarRegularity
open MeasureTheory
open scoped ENNReal
/-- For measurable `h, f : ℝ^d → ℂ`,
`∫ (∫ |h(x - y)| |f(y)| dy)² dx ≤ (∫ |h|²) (∫ |f|)²`. -/
theorem lintegral_sq_lintegral_mul_le {d : ℕ} (h f : EuclideanSpace ℝ (Fin d) → ℂ)
(hh : AEStronglyMeasurable h volume) (hf : AEStronglyMeasurable f volume) :
∫⁻ x, (∫⁻ y, ‖h (x - y)‖ₑ * ‖f y‖ₑ) ^ 2
≤ (∫⁻ x, ‖h x‖ₑ ^ 2) * (∫⁻ y, ‖f y‖ₑ) ^ 2 := by sorry
end DbarRegularity
Source
E. M. Stein, Singular Integrals and Differentiability Properties of Functions, Princeton University Press (1970), Chapter II, § 2.2, p. 29