Documentation

LeanPool.HighContrastHomogenization.Support.CoarseGraining.BlockFormalism.EllipticBounds

Coarse-graining support: Support.CoarseGraining.BlockFormalism.EllipticBounds #

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

Block formalism -- symmetric and elliptic matrix bounds #

Symmetric block-matrix identities, symmPartInv lower / upper bound and positivity under IsEllipticMatrix, blockMatrixOfCoeff_half_quadratic_ge_vecDot, blockMatrixOfCoeff_quadratic positivity / lower / upper / plainUpperBound, coercivity, and the image bounds used by the response functional.

theorem HCPolySupport.symmPart_inv_nonneg_of_isEllipticMatrix {d : ℕ} {lam Lam : ℝ} {A : Mat d} (hA : IsEllipticMatrix lam Lam A) (ξ : Vec d) :
theorem HCPolySupport.lowerBound_symmPartInv_of_isEllipticMatrix {d : ℕ} {lam Lam : ℝ} {A : Mat d} (hA : IsEllipticMatrix lam Lam A) (ξ : Vec d) :
lam * (Lam⁻¹ * Lam⁻¹) * vecNormSq ξ ≤ vecDot ξ (matVecMul (symmPart A)⁻¹ ξ)
theorem HCPolySupport.blockMatrixOfCoeff_image_plainUpperBound_of_isEllipticMatrix {d : ℕ} {lam Lam : ℝ} {A : Mat d} (hA : IsEllipticMatrix lam Lam A) (p q : Vec d) :
blockVecDot (blockMatVecMul (blockMatrixOfCoeff A) (p, q)) (blockMatVecMul (blockMatrixOfCoeff A) (p, q)) ≤ (2 * Lam ^ 2 + 2 * (2 * Lam ^ 2 + 1) * (lam⁻¹ * lam⁻¹) * Lam ^ 2) * vecNormSq p + 2 * (2 * Lam ^ 2 + 1) * (lam⁻¹ * lam⁻¹) * vecNormSq q
noncomputable def HCPolySupport.blockMatrixOfCoeffNormSqBound (lam Lam : ℝ) :

An explicit squared operator-norm bound for the block matrix induced by a coefficient matrix with ellipticity parameters lam and Lam.

Equations
Instances For