Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section4.Provider.Regularity.StepSevenMean

StepSevenMean #

Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.

t.regularity Step 7d: the hmean producer, and the chain without hmean #

The gap this module closes #

  oscTrunc ≤ Cmean · ( (3^{Q.scale})^{-1} · osc_{L²}(Q; u) )

as a hypothesis of the narrowed Step-7d chain. Producers do exist, in two steps, at Cmean = √(3^d):

This module composes them (stepSevenHmean_of_sandwich) and proves the narrowed chain with hmean gone (exists_stepSevenEnd_chain_of_lambda_hmeanFree).

The carriers, and where the two constants come from #

At the scale normalization Q = originCube d k (k = m'-1 at the consumption site) with the sample translated by c:

  oscTrunc = 3^{-k}‖u - (u)_{(z+□_k)∩□_m}‖_{L²((z+□_k)∩□_m)}     (the tex's LHS)
           ≤ 3^{d/2} · 3^{-k}‖u - (u)_{c+□_k}‖_{L²(c+□_k)}        (sandwich)
           ≤ 3^{d/2} · 3^{-k}·osc_{L²}(□_k; u(·+c)) .              (cost 1)

So Cmean = √(3^d) and nothing else enters.

What is still the caller's #

The five integrability binders of the two proved atoms (w and w² on both windows, and the mean-subtracted square on the sandwich cube) and the sandwich inclusion hST — the latter is StepSevenSelectionData.upper_sandwich_in at k = m'-1, so it is free at the selection. The three measure-theoretic side conditions of the sandwich cube (measurability, positive volume, finite volume) are discharged here rather than carried: a translate of the open cube is open and has volume 3^{dk}.

References #

1. The sandwich cube's measure-theoretic data, discharged #

2. The hmean producer #

theorem Algsuperdiff.Section4.Provider.Regularity.stepSevenHmean_of_sandwich {d : ℕ} {m k : ℤ} {z c : HCPolySupport.Vec d} {w : HCPolySupport.Vec d → ℝ} (v : HCPolySupport.H1Function (HCPolySupport.openCubeSet (HCPolySupport.originCube d k))) (hv : ∀ (y : HCPolySupport.Vec d), v.toFun y = w (y + c)) (hz : z ∈ HCPolySupport.openCubeSet (HCPolySupport.originCube d m)) (hkm : k ≤ m) (hST : ExcessDecay.truncatedWindow z m k ⊆ (fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) (hwS : MeasureTheory.IntegrableOn w (ExcessDecay.truncatedWindow z m k) MeasureTheory.volume) (hwS2 : MeasureTheory.IntegrableOn (fun (x : HCPolySupport.Vec d) => w x ^ 2) (ExcessDecay.truncatedWindow z m k) MeasureTheory.volume) (hwT : MeasureTheory.IntegrableOn w ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) MeasureTheory.volume) (hwT2 : MeasureTheory.IntegrableOn (fun (x : HCPolySupport.Vec d) => w x ^ 2) ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) MeasureTheory.volume) (hint : MeasureTheory.IntegrableOn (fun (x : HCPolySupport.Vec d) => (w x - HCPolySupport.volumeAverage ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) w) ^ 2) ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) MeasureTheory.volume) :
  3^{-k}‖u - (u)_{(z+□_k)∩□_m}‖_{L²((z+□_k)∩□_m)}
      ≤ √(3^d) · ( 3^{-k} · osc_{L²}(□_k; u(·+c)) ) ,

which is the hmean slot of the Step-7d chain at Cmean = √(3^d) and Q = originCube d k. Two proved atoms composed, nothing else.

The Step-7d transported oscillation is nonnegative — the chain's hoscTrunc at the concrete carrier, so the caller need not supply it.

3. The narrowed chain, with hmean gone #

theorem Algsuperdiff.Section4.Provider.Regularity.exists_stepSevenEnd_chain_of_lambda_hmeanFree (d : ℕ) [NeZero d] :
∃ (C : ℝ), 0 < C ∧ ∀ {alpha : ℝ} {n m k : ℤ} {a : HCPolySupport.Book.Ch03.CoeffFamily d} {g : HCPolySupport.Vec d → HCPolySupport.Vec d} {z c : HCPolySupport.Vec d} {w : HCPolySupport.Vec d → ℝ} (u : HCPolySupport.Book.Ch03.ForcedCubeSolution (HCPolySupport.originCube d k) a g) {Cg Clam Ccmp Kg Kd Ks Ctr shomNp shomMp shomM gradLoc gradM dataG dataM W G H : ℝ}, (∀ (y : HCPolySupport.Vec d), u.toH1.toFun y = w (y + c)) → z ∈ HCPolySupport.openCubeSet (HCPolySupport.originCube d m) → k ≤ m → ExcessDecay.truncatedWindow z m k ⊆ (fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k) → MeasureTheory.IntegrableOn w (ExcessDecay.truncatedWindow z m k) MeasureTheory.volume → MeasureTheory.IntegrableOn (fun (x : HCPolySupport.Vec d) => w x ^ 2) (ExcessDecay.truncatedWindow z m k) MeasureTheory.volume → MeasureTheory.IntegrableOn w ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) MeasureTheory.volume → MeasureTheory.IntegrableOn (fun (x : HCPolySupport.Vec d) => w x ^ 2) ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) MeasureTheory.volume → MeasureTheory.IntegrableOn (fun (x : HCPolySupport.Vec d) => (w x - HCPolySupport.volumeAverage ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) w) ^ 2) ((fun (y : HCPolySupport.Vec d) => c + y) '' HCPolySupport.openCubeSet (HCPolySupport.originCube d k)) MeasureTheory.volume → 0 ≤ Cg → 0 ≤ Clam → 0 ≤ Ks → 0 ≤ Ccmp → 0 ≤ shomMp → 0 < shomM → 0 ≤ gradM → 0 ≤ dataG → 0 ≤ W → 0 ≤ G → 0 ≤ H → HCPolySupport.Book.Ch03.ForceBesovRegularity (HCPolySupport.originCube d k) stepSevenCgS g → stepSevenCgLamInv (HCPolySupport.originCube d k) a stepSevenCgS ≤ Clam * shomMp → HCPolySupport.Book.Ch03.forcedSolutionEnergyNorm (HCPolySupport.originCube d k) a u ≤ Kg * gradM → HCPolySupport.Book.Ch03.scaleNormalizedPositiveBesovVectorSeminormTwo (HCPolySupport.originCube d k) stepSevenCgS g ≤ Kd * dataG → shomMp ≤ Ks * shomM⁻¹ → √Ks * Kg ≤ Ctr * Real.rpow 3 (1 / 4 * stepSixExponent alpha n m) → Ks * Kd ≤ Ctr * Real.rpow 3 (1 / 4 * stepSixExponent alpha n m) → shomNp ≤ Ccmp * shomM → gradLoc ≤ Cg * √shomNp * Real.rpow 3 (3 / 4 * stepSixExponent alpha n m) * ((HCPolySupport.cubeScaleFactor (HCPolySupport.originCube d k))⁻¹ * Support.normalizedL2On (ExcessDecay.truncatedWindow z m k) fun (x : HCPolySupport.Vec d) => w x - HCPolySupport.volumeAverage (ExcessDecay.truncatedWindow z m k) w) + Cg * Real.rpow 3 (3 / 4 * stepSixExponent alpha n m) * (√shomNp * (W * (shomM⁻¹ * G + H)) + dataM) → gradLoc ≤ Cg * √Ccmp * (√(3 ^ d) * stepSevenEmbeddingConst d * stepSevenBridgeConst stepSevenCgS * (C * 64 * (√Clam + Clam)) * Ctr) * Real.rpow 3 (stepSixExponent alpha n m) * (gradM + √shomM⁻¹ * dataG) + Cg * Real.rpow 3 (3 / 4 * stepSixExponent alpha n m) * (√Ccmp * (W * ((√shomM)⁻¹ * G + √shomM * H)) + dataM)

Step 7d end to end with hcg, hembed, hbridge AND hmean discharged.

exists_stepSevenEnd_chain_of_lambda_gen fed by stepSevenHmean_of_sandwich, at Cmean = √(3^d) and the concrete transported oscillation

  oscTrunc = 3^{-k}‖u - (u)_{(z+□_k)∩□_m}‖_{L²((z+□_k)∩□_m)} .

hoscTrunc is discharged too (it is a product of nonnegatives at the concrete carrier). The remaining conditional inputs are hlambda (e.lambda.stability.applied) and the Step-7c gradient display hgrad; everything else is the caller's geometry, integrability, or the printed nonnegativity data.