Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3.27 — Gagliardo / Loomis–Whitney product inequality (3.10)

Proved
HunterPDE.Sobolev.product_inequality

by mikedeng1 · Sep 28, 2026 · Mathlib 0df444a (Lean v4.33.1)

inequalitiesp2o-batch-books5p2o-gran-per-chapterp2o-plan-bookp2o-v1pdesobolev-spaces

Let n≥2n \ge 2n≥2. For x=(x1,…,xn)∈Rnx = (x_1, \dots, x_n) \in \mathbb{R}^nx=(x1​,…,xn​)∈Rn and 1≤i≤n1 \le i \le n1≤i≤n let xi′=(x1,…,x^i,…,xn)∈Rn−1x_i' = (x_1, \dots, \hat{x}_i, \dots, x_n) \in \mathbb{R}^{n-1}xi′​=(x1​,…,x^i​,…,xn​)∈Rn−1 be xxx with its iiith coordinate omitted. Suppose g1,…,gn∈Cc∞(Rn−1)g_1, \dots, g_n \in C_c^\infty(\mathbb{R}^{n-1})g1​,…,gn​∈Cc∞​(Rn−1) are nonnegative and define g(x)=∏i=1ngi(xi′)g(x) = \prod_{i=1}^n g_i(x_i')g(x)=∏i=1n​gi​(xi′​). Then

∫Rng dx≤∏i=1n∥gi∥n−1,(3.10)\int_{\mathbb{R}^n} g \, dx \le \prod_{i=1}^n \|g_i\|_{n-1}, \qquad (3.10)∫Rn​gdx≤i=1∏n​∥gi​∥n−1​,(3.10)

where ∥gi∥n−1\|g_i\|_{n-1}∥gi​∥n−1​ is the Ln−1(Rn−1)L^{n-1}(\mathbb{R}^{n-1})Ln−1(Rn−1) norm. For n=2n = 2n=2 this is Fubini's theorem; in general it estimates the L1L^1L1 norm of a function of xxx by the Ln−1L^{n-1}Ln−1 norms of nnn functions of the xi′x_i'xi′​ whose product bounds it. It is the multilinear step in the proof of Theorem 3.28.

Formalization Note. nnn is written m+1m + 1m+1 with m≥1m \ge 1m≥1, so Rn−1\mathbb{R}^{n-1}Rn−1 is EuclideanSpace ℝ (Fin m) without natural-number subtraction; indices are 0-based, i : Fin (m + 1), and xi′x_i'xi′​ is Fin.removeNth i applied to the coordinates of xxx. The left side is the Lebesgue integral of the nonnegative function ggg in [0,∞][0, \infty][0,∞]; the norms are eLpNorm (g i) m.

Preamble
import Mathlib

open MeasureTheory
Formal statement
namespace HunterPDE.Sobolev

/-- Theorem 3.27 of Hunter, *Notes on PDEs* (revised 6/18/2014), p. 61 (Gagliardo / Loomis–Whitney
inequality (3.10)). Write `n = m + 1` with `n ≥ 2` (i.e. `m ≥ 1`). Let `gᵢ ∈ C_c^∞(ℝ^{n−1})`,
`i = 1, …, n` (here `i : Fin (m + 1)`, 0-based), be nonnegative, and let
`g(x) = ∏ᵢ gᵢ(x′ᵢ)`, where `x′ᵢ ∈ ℝ^{n−1}` is `x` with its `i`th coordinate omitted
(`Fin.removeNth i`). Then `∫ g dx ≤ ∏ᵢ ‖gᵢ‖_{n−1}`, the norms being `L^{n−1}(ℝ^{n−1})` norms.
The left side is the Lebesgue integral of the nonnegative function `g`. -/
theorem product_inequality {m : ℕ} (hm : 1 ≤ m)
    (g : Fin (m + 1) → EuclideanSpace ℝ (Fin m) → ℝ)
    (hsmooth : ∀ i, ContDiff ℝ ((⊤ : ℕ∞) : WithTop ℕ∞) (g i))
    (hcpt : ∀ i, HasCompactSupport (g i))
    (hnonneg : ∀ i y, 0 ≤ g i y) :
    ∫⁻ x : EuclideanSpace ℝ (Fin (m + 1)),
        ∏ i, ENNReal.ofReal (g i (WithLp.toLp 2 (Fin.removeNth i (WithLp.ofLp x)))) ≤
      ∏ i, eLpNorm (g i) (m : ENNReal) volume := by sorry

end HunterPDE.Sobolev
Source
Hunter, Notes on Partial Differential Equations (revised 6/18/2014), p. 61, Theorem 3.27
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Let m≥1m \ge 1m≥1 be a natural number and write points of Rm+1\mathbb{R}^{m+1}Rm+1 as x=(x0,…,xm)x = (x_0, \dots, x_m)x=(x0​,…,xm​). For each i∈{0,…,m}i \in \{0, \dots, m\}i∈{0,…,m}, let xi′∈Rmx'_i \in \mathbb{R}^mxi′​∈Rm be xxx with its iii-th coordinate deleted, keeping the remaining coordinates in their original order.

Let g0,…,gm:Rm→Rg_0, \dots, g_m : \mathbb{R}^m \to \mathbb{R}g0​,…,gm​:Rm→R be m+1m+1m+1 functions, each of which is:

  • C∞C^\inftyC∞;
  • compactly supported;
  • nonnegative everywhere.

Then

∫Rm+1∏i=0mgi(xi′) dx  ≤  ∏i=0m∥gi∥Lm(Rm).\int_{\mathbb{R}^{m+1}} \prod_{i=0}^{m} g_i(x'_i)\,dx \;\le\; \prod_{i=0}^{m} \|g_i\|_{L^m(\mathbb{R}^m)} .∫Rm+1​i=0∏m​gi​(xi′​)dx≤i=0∏m​∥gi​∥Lm(Rm)​.

The terms in this inequality are as follows:

  • The left side is the Lebesgue integral of a nonnegative function, valued in [0,∞][0,\infty][0,∞]. Each factor is converted to an extended nonnegative real, which is exact here since gi≥0g_i \ge 0gi​≥0.
  • On the right, ∥h∥Lm(Rm)=(∫Rm∣h∣m)1/m\|h\|_{L^m(\mathbb{R}^m)} = \left(\int_{\mathbb{R}^m} |h|^m\right)^{1/m}∥h∥Lm(Rm)​=(∫Rm​∣h∣m)1/m with respect to Lebesgue measure on Rm\mathbb{R}^mRm.

Degenerate cases. The value m=0m = 0m=0 is excluded by hypothesis. For m=1m = 1m=1 the statement reads

∬g0(x1) g1(x0) dx0 dx1≤∥g0∥1 ∥g1∥1.\iint g_0(x_1)\,g_1(x_0)\,dx_0\,dx_1 \le \|g_0\|_{1}\,\|g_1\|_{1}.∬g0​(x1​)g1​(x0​)dx0​dx1​≤∥g0​∥1​∥g1​∥1​.

If some gi≡0g_i \equiv 0gi​≡0, both sides are 000. The right side is finite, because each gig_igi​ is continuous with compact support. The hypotheses are satisfiable, for example by all gi=0g_i = 0gi​=0.

Human review
  • Endorsed by Shuze Chen · Oct 2, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 2, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me