Coarse-graining support: Support.Book.Ch02.MultiscaleEllipticity #
Imported from the Apache-2.0 CoarseGraining development at commit
c7ddd76c08ade64fed1b8d2ca51be14dfee8deb4.
Chapter 2.5 Multiscale Ellipticity Constants #
This file locks the public definitions from Section 2.5 of the notes. The
multiscale quantities are deliberately built from the Chapter 2 CoeffOn
interface on open cube domains, and the compatibility of a coefficient field
across nested cubes is recorded almost everywhere.
The open realization of a triadic cube is a nonempty public Chapter 2 domain.
The public Chapter 2 domain associated with an open triadic cube.
Equations
- HCPolySupport.Book.Ch02.cubeDomain Q = { carrier := HCPolySupport.openCubeSet Q, isDomain := ⋯, nonempty := ⋯ }
Instances For
A coefficient field on the triadic cube hierarchy.
For each open triadic cube it provides a CoeffOn object, hence ellipticity and
measurability are a.e. on that cube. The restrictsTo_of_subset field says
that the representatives are compatible across nested cubes only modulo null
sets; this is the public replacement for old representative-level cube
restrictions.
- coeffOn (Q : TriadicCube d) : CoeffOn (cubeDomain Q)
The measurable, almost-everywhere elliptic coefficient object assigned to each open triadic cube.
- restrictsTo_of_subset {Q R : TriadicCube d} : openCubeSet R ⊆ openCubeSet Q → (self.coeffOn Q).RestrictsTo (self.coeffOn R)
Instances For
Two triadic coefficient families are equal when their cube representatives agree a.e. on every open triadic cube.
Instances For
Compatibility of a triadic coefficient family with a descendant cube, expressed a.e. on the descendant open cube.
Exponents used in the multiscale ellipticity constants: finite q and the
endpoint q = infinity.
- finite (value : ℝ) : MultiscaleExponent
- infinity : MultiscaleExponent
Instances For
Equations
- One or more equations did not get rendered due to their size.
- HCPolySupport.Book.Ch02.instDecidableEqMultiscaleExponent.decEq (HCPolySupport.Book.Ch02.MultiscaleExponent.finite value) HCPolySupport.Book.Ch02.MultiscaleExponent.infinity = isFalse ⋯
- HCPolySupport.Book.Ch02.instDecidableEqMultiscaleExponent.decEq HCPolySupport.Book.Ch02.MultiscaleExponent.infinity (HCPolySupport.Book.Ch02.MultiscaleExponent.finite value) = isFalse ⋯
- HCPolySupport.Book.Ch02.instDecidableEqMultiscaleExponent.decEq HCPolySupport.Book.Ch02.MultiscaleExponent.infinity HCPolySupport.Book.Ch02.MultiscaleExponent.infinity = isTrue ⋯
Instances For
Admissible exponents in the public Sec. 2.5 definitions: finite q ≥ 1
or the endpoint q = infinity.
Equations
Instances For
Legacy Frobenius-style squared matrix norm retained for compatibility with older deterministic infrastructure.
Equations
- HCPolySupport.Book.Ch02.matrixNormSq A = ∑ i : Fin d, ∑ j : Fin d, A i j ^ 2
Instances For
Euclidean/L2 operator norm used for |b| and |sigma_*^{-1}| in the
public Sec. 2.5 definitions.
Equations
Instances For
Supremum of a real-valued function over a finite set.
The old deterministic files use the same sSup convention, which leaves the
definition total even when the finite set is empty.
Equations
- HCPolySupport.Book.Ch02.finsetSupReal s f = sSup (f '' ↑s)
Instances For
The scale factor 3^{2s(m-k)} appearing in one-cube descendant bounds.
Equations
Instances For
One-cube norm |b(Q; a)|, where b is the canonical public coarse matrix
on the open cube.
Equations
Instances For
One-cube norm |sigma_*^{-1}(Q; a)|, where sigma_*^{-1} is the canonical
public coarse matrix on the open cube.
Equations
Instances For
The maximum of |b(R; a)| over descendants of Q at scale k.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The maximum of |sigma_*^{-1}(R; a)| over descendants of Q at scale
k.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Geometric normalization c_{s,q} = 1 - 3^{-s q} for finite exponents.
Instances For
The finite-exponent geometric weight
c_{s,q} 3^{-s q n}.
Equations
- HCPolySupport.Book.Ch02.geometricWeight s q n = HCPolySupport.Book.Ch02.geometricDiscount s q * Real.rpow 3 (-s * q * ↑n)
Instances For
Finite-exponent coarse-grained upper ellipticity
\Lambda_{s,q}(Q; a).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Finite-exponent coarse-grained lower ellipticity
\lambda_{s,q}(Q; a).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Endpoint coarse-grained upper ellipticity
\Lambda_{s,\infty}(Q; a).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Endpoint coarse-grained lower ellipticity
\lambda_{s,\infty}(Q; a).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Coarse-grained upper ellipticity for finite q and q = infinity.
Equations
Instances For
Coarse-grained lower ellipticity for finite q and q = infinity.
Equations
Instances For
The maximum of \Lambda_{s,q}(R; a) over descendants of Q at scale
k.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The maximum of \lambda_{s,q}(R; a)^{-1} over descendants of Q at scale
k.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The default finite-exponent convention q = 1 for \Lambda_s.
Equations
Instances For
The default finite-exponent convention q = 1 for \lambda_s.
Equations
Instances For
Coarse-grained ellipticity ratio
\Theta_{s,t}(Q; a) = \Lambda_{s,1}(Q; a) / \lambda_{t,1}(Q; a).