High-contrast homogenization: #
Provider.PolynomialHomogenization.Root.CorrectorComposition.CorrectorEnergyOnCube
Imported from the Apache-2.0 HighContrastHomogenization development at commit
7a13dbcd8d6609264a713373f5c69ceeac870472.
The corrector's own energy, in the representative form the terminal uses #
The exists_scalarIdentityGoodTailJointWeightedGradientBoundConstant
bounds the affine-plus-corrector energy on every cube above the good-tail start
by B * euclideanNorm e, with B law-free. It is stated on the local H¹
object finiteAffineCorrectionJointLocalH1 a hCauchy e q; the large-scale C¹ slope approximation
terminal
speaks of fun y ↦ e + (Phi e).globalGradientRepresentative y.
The two agree a.e. on the cube — finiteAffineCorrectionJointLocalH1 is by
definition the affine boundary function plus the carrier's local H¹
function, and globalGradientRepresentative_ae_eq_localH1Gradient identifies
the representative with that local gradient. This module performs that one
weightedGradNorm_congr_ae, for an arbitrary joint-local-equation carrier
Phi (pinned to the canonical limit by
isFiniteAffineCorrectionJointLocalLimit_iff_eq).
Nothing here is an estimate.
The corrector energy bound, in representative form. Above the
good-tail start the affine-plus-corrector energy on every origin cube is at
most B * euclideanNorm e, with B depending only on (d, s).