Documentation

LeanPool.HighContrastHomogenization.Support.Book.Ch03.Theorems.CoarseCaccioppoliRHS.EnergySplit

Coarse-graining support: Support.Book.Ch03.Theorems.CoarseCaccioppoliRHS.EnergySplit #

Imported from the Apache-2.0 CoarseGraining development at commit c7ddd76c08ade64fed1b8d2ca51be14dfee8deb4.

Coarse Caccioppoli with RHS Energy Split #

This file is split mechanically out of CoarseCaccioppoliRHS.lean.

Audit tag #

Claim: split forced boundary Caccioppoli energy into the homogeneous part, zero-trace corrector part, and cross terms using the public coefficient representative.

Downstream target: CoarseCaccioppoliRHS/Prefactors.lean and CoarseCaccioppoliRHS/FinalBounds.lean. This file should contain energy identities and inequalities only.

On subsets of the open cube, public localized energy can be evaluated using the pointwise deterministic coefficient representative.

theorem HCPolySupport.Book.Ch03.volumeAverage_coefficientEnergyDensity_le_two_mul_add_of_ae_eq_add {d : ℕ} {V : Set (Vec d)} {A : CoeffField d} {lam Lam : ℝ} {F G H : Vec d → Vec d} (hEll : IsEllipticFieldOn lam Lam V A) (hF : MemVectorL2 V F) (hG : MemVectorL2 V G) (hH : MemVectorL2 V H) (hFGH : F =ᵐ[volumeMeasureOn V] fun (y : Vec d) => G y + H y) :

Coefficient-energy triangle inequality for a decomposition F = G + H over an arbitrary measurable set.

theorem HCPolySupport.Book.Ch03.normalizedSetAverage_caccioppoliCoreSet_le_eighteen_pow_mul_cubeAverage {d : ℕ} (Q : TriadicCube d) {x : Vec d} {energy : Vec d → ℝ} (hx : x ∈ openCubeSet Q) (henergy_nonneg : ∀ y ∈ cubeSet Q, 0 ≤ energy y) (henergy_int : MeasureTheory.IntegrableOn energy (cubeSet Q) MeasureTheory.volume) :

A core average is controlled by the parent cube average with the dimension-only volume ratio 18^d.

The zero-trace corrector's localized core energy is controlled by its parent cube coefficient energy with only the geometric 18^d loss.

theorem HCPolySupport.Book.Ch03.zeroDirichletEnergyWithRHSRHS_nonneg {d : ℕ} {C : ℝ} (hC_nonneg : 0 ≤ C) {Q : TriadicCube d} {a : CoeffFamily d} {t : ℝ} {g : Vec d → Vec d} (ht : 0 < t) (hg : ForceBesovRegularity Q (2 * t) g) :

The public zero-Dirichlet RHS is nonnegative under the force regularity hypothesis.

theorem HCPolySupport.Book.Ch03.corrector_parentEnergy_le_zeroDirichletRHS_squared {d : ℕ} [NeZero d] {C : ℝ} (hC_nonneg : 0 ≤ C) (hC_zero : √(250 + 2 * √15000 * √2) * (↑d * Real.rpow 3 (↑d + 1)) ≤ C) {Q : TriadicCube d} {a : CoeffFamily d} {t : ℝ} {g : Vec d → Vec d} (ρ : ZeroTraceDirichletCorrectorData Q (publicCoeffField Q a) g) (ht : 0 < t) (ht_lt : t < 1 / 2) (hg : ForceBesovRegularity Q (2 * t) g) :

Squared form of the zero-trace corrector energy estimate, tuned to the t-notation used by the boundary Caccioppoli RHS theorem.

theorem HCPolySupport.Book.Ch03.boundaryForcedCaccioppoliCorrector_coreEnergy_le_eighteen_pow_mul_zeroDirichletEnergyWithRHSRHS_sq {d : ℕ} [NeZero d] {C : ℝ} (hC_nonneg : 0 ≤ C) (hC_zero : √(250 + 2 * √15000 * √2) * (↑d * Real.rpow 3 (↑d + 1)) ≤ C) {Q : TriadicCube d} {a : CoeffFamily d} {t : ℝ} {x : Vec d} {g : Vec d → Vec d} (ρ : ZeroTraceDirichletCorrectorData Q (publicCoeffField Q a) g) (ht : 0 < t) (ht_lt : t < 1 / 2) (hx : x ∈ openCubeSet Q) (hg : ForceBesovRegularity Q (2 * t) g) :

The localized corrector core is controlled directly by the square of the public zero-Dirichlet RHS, with only the geometric 18^d loss.

Apply the proved homogeneous boundary Caccioppoli theorem to the harmonic remainder produced from a forced datum.