High-contrast homogenization: Provider.Regularity.GoodTail #
Imported from the Apache-2.0 HighContrastHomogenization development at commit
7a13dbcd8d6609264a713373f5c69ceeac870472.
Summable weak-error tails on centered cubes #
The normalized large-scale regularity argument uses the finite sum of the Chapter 2 homogenization errors on centered Euclidean triadic cubes. This module records the finite-interval and infinite-tail predicates for the identity comparison matrix.
The multiscale weak error on the centered Euclidean triadic cube at scale
k, with identity comparison matrix and the p = infinity, q = 2
exponents.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The identity weak error is nonnegative.
The finite l¹ weak-error row between two integer scales.
Equations
- HCPolySupport.HighContrast.ScalarIdentityGoodTailOnInterval a s δ n m = (∑ k ∈ Finset.Icc n m, HCPolySupport.HighContrast.scalarIdentityWeakError a s k ≤ δ)
Instances For
The l¹ weak-error row is bounded on every finite interval beginning at
n.
Equations
- HCPolySupport.HighContrast.ScalarIdentityGoodTail a s δ n = ∀ (m : ℤ), n ≤ m → HCPolySupport.HighContrast.ScalarIdentityGoodTailOnInterval a s δ n m
Instances For
Increasing the tolerance preserves a finite good-tail bound.
A good tail supplies its bound on every admissible finite interval.
Increasing the tolerance preserves an infinite good-tail bound.
Discarding initial scales preserves a finite good-tail bound.
Discarding initial scales preserves an infinite good-tail bound.
Every weak error in a good tail is bounded by its tolerance.