High-contrast homogenization: Provider.Regularity.CorrectorRealRadiusNegOne #
Imported from the Apache-2.0 HighContrastHomogenization development at commit
7a13dbcd8d6609264a713373f5c69ceeac870472.
Quotient-safe negative-one restriction to centered real cubes #
This module restricts a Hilbert-vector L² class from an outer domain to a
measurable subdomain and compares the normalized negative-one dual norms. A
compactly supported test on the inner domain is also a test on the outer
domain. Exact support identities for its value and gradient energies produce
the inner-to-outer volume ratio in the squared test norm, hence the square root
of the reciprocal ratio in the dual norm.
The final specialization uses the canonical real-radius origin cube from
CorrectorRealRadiusGeometry. No corrector decay or tail estimate is used.
Restriction of local Hilbert-vector classes #
The real linear restriction map for Hilbert-vector L² classes along the inclusion
U ⊆ V.
Equations
- HCPolySupport.HighContrast.localHilbertVectorL2RestrictLinear hUV = { toFun := fun (F : ↥(HCPolySupport.HilbertVectorL2 V)) => MeasureTheory.MemLp.toLp ↑↑F ⋯, map_add' := ⋯, map_smul' := ⋯ }
Instances For
Continuous restriction of a Hilbert-vector L² class to an arbitrary
set contained in its original domain.
Equations
Instances For
Local Hilbert-vector restriction agrees almost everywhere with the outer representative on the inner domain.
Compact support and normalized test energies #
A local vector test remains a local test after enlarging its domain.
Enlarging the domain decreases the normalized squared H¹ test norm
by at least the inner-to-outer volume ratio for a test supported in the inner
domain.
The normalized negative-one restriction #
A local test whose squared H¹ norm is at most K pairs by at most
sqrt K times the negative-one norm.
A raw local L² field restricted from V to U has normalized
negative-one norm at most the square root of the outer-to-inner volume ratio
times its outer norm.
Quotient-safe restriction of a local Hilbert-vector class costs the same square-root volume ratio in the normalized negative-one norm.