Nonnegative Minkowski integral inequality with finite mixed norm
ProvedMeasureTheory.eLpNorm_lintegral_le_of_lt_topinequalitiesintegrationlp-spacesmeasure-theory
Let and be measurable spaces with s-finite measures, let be a real exponent, and let be almost everywhere measurable for the product measure. Suppose
Then
All integrals are nonnegative extended integrals. There is no pointwise finiteness or integrability assumption on the sections. This is the nonnegative mixed-norm form of Minkowski's integral inequality, usable for measurable kernels independently of signed Bochner integration. The endpoint and the infinite-bound case are excluded because they follow directly from Tonelli and the extended order, respectively.
Preamble
import Mathlib.MeasureTheory.Integral.Prod import Mathlib.MeasureTheory.Function.LpSeminorm.TriangleInequality open MeasureTheory
Formal statement
theorem MeasureTheory.eLpNorm_lintegral_le_of_lt_top
{α β : Type*} [MeasurableSpace α] [MeasurableSpace β]
(μ : Measure α) (ν : Measure β) [SFinite μ] [SFinite ν]
{q : ℝ} (hq : 1 < q) (G : α → β → ENNReal)
(hG : AEMeasurable (Function.uncurry G) (μ.prod ν))
(hbound : (∫⁻ y, eLpNorm (fun x => G x y) (ENNReal.ofReal q) μ ∂ν) < ⊤) :
eLpNorm (fun x => ∫⁻ y, G x y ∂ν) (ENNReal.ofReal q) μ ≤
∫⁻ y, eLpNorm (fun x => G x y) (ENNReal.ofReal q) μ ∂ν := by sorrySource
General nonnegative mixed-norm form of Minkowski integral inequality, derived using Holder and Tonelli or weighted Jensen. John K. Hunter, Notes on Partial Differential Equations, revised 6/18/2014, https://www.math.ucdavis.edu/~hunter/pdes/pde_notes.pdf, Section 1.4, Theorems 1.7 (Jensen) and 1.9 (Holder), printed pp. 4-5, and Appendix 3.C, printed pp. 86-89. This s-finite two-measure interface is a derived generalization, not a verbatim statement in Hunter.