Documentation

LeanPool.HighContrastHomogenization.Provider.Response.PreYoungAdjointGradientPhysicalProjectionLimit

High-contrast homogenization: Provider.Response.PreYoungAdjointGradientPhysicalProjectionLimit #

Imported from the Apache-2.0 HighContrastHomogenization development at commit 7a13dbcd8d6609264a713373f5c69ceeac870472.

The physical adjoint gradient projection limit #

Finite projected adjoint gradient pairings converge to the physical cutoff oscillation. Their common annealed row bound passes to the limit by Fatou's lemma.

theorem HCPolySupport.HighContrast.Response.tendsto_adjoint_gradient_projected_oscillation {d : ℕ} [NeZero d] {q : Mat d} (hq : Matrix.PosDef q) {s t : ℤ} (hst : s ≤ t) (g : Mat d) (hg : IsSkewMat g) (a : CoeffSpace d) (p r Qcen : Vec d) :
Filter.Tendsto (fun (n : ℕ) => adjointGradientProjectedOscillation hq s t g hg p r Qcen (n + 1) a) Filter.atTop (nhds (adjointGradientPhysicalOscillation hq s t g hg p r Qcen a))

Finite-depth projected adjoint gradient pairings converge pointwise to the physical pairing.

theorem HCPolySupport.HighContrast.Response.adjoint_gradient_physical_parent_pairing_eq_coordinate_sum {d : ℕ} [NeZero d] {q : Mat d} (hq : Matrix.PosDef q) {s t : ℤ} (hst : s ≤ t) {z : Fin d → ℤ} (hz : z ∈ alignedIndex q s t) (g : Mat d) (hg : IsSkewMat g) (a : CoeffSpace d) (p r Qcen : Vec d) :
have U := adaptedCellAt q s z; have cut := adaptedPreYoungCutoff q hq t; have X := diagonalWeakAdjointState hq t (a.subSkew g hg) p r; have R := translateCube z (originCube d s); have phi := fun (y : Vec d) => cut (matVecMul q y); have f := fun (y : Vec d) => phi y - cubeAverage R phi; have G := fun (y : Vec d) => vecDot Qcen (X (matVecMul q y)).1; cubeBesovPairing R f G = ∑ i : Fin d, Qcen i * volumeAverage U fun (x : Vec d) => (cut x - volumeAverage U cut) * (X x).1 i

The physical adjoint gradient pairing on one parent cell is its coordinate expansion.

A finite adjoint weak quantity makes the physical gradient cutoff oscillation integrable.