LocalizationFluctuationBackgroundClass #
Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.
The Step-2 background's L^2 class is measurable in the sample #
ABK26, l.approximate.recurrence.formula, Step 2, e.Pz.def, e.Fz.def, the
background display.
The obligation this module discharges #
LocalizationFluctuationJointBackground reduced the whole hcomp/hbg
measurability front of
LocalizationFluctuationMuMeasCell.integrable_meshFluctuationCellIntegrand' to
the single proposition
omega |-> [ meshCellBackground M n h Kc e e' wD wN R omega ]
is measurable into L^2(z + cu_n; R^{2d}). This module proves it.
Method: the cube average cancels #
The Step-2 background is P_z + bfF_z, and both summands carry the cube average
(grad w)_{z+cu_n}: P_z with a + sign (e.Pz.def) and bfF_z with a -
sign (e.Fz.def). In the sum the average cancels, and
PrincipalResponsePz.principalPz_add_principalFz records exactly that:
P_z + bfF_z(x) = bfAhom^{-1/2} ( e' + grad w_D(x); e + grad w_N(x) + shom^{-1} h e' ) .
So the background is LocalizationActualBackground.localizationBackground,
which does not depend on the localization cube at all, and no measurability
of any cube average is needed.
The class of that field is then built on the localization cube cu_K, where
the proved measurability lives, and restricted to the mesoscopic cell:
principalLoad_addsplits off the constantbfAhom^{-1/2}(e'; e), whose class is the continuous linearblockVecToHilbertBlockL2Const(LocalizationFluctuationJointComponents.toHilbertBlockL2OfBlockField_const), and adding it back translates the class by that constant, which preserves measurability (LocalizationFluctuationJointBackground.measurable_toHilbertBlockL2OfBlockField_add_const);- what is left is the doubled class of the pair of scaled corrector gradients,
assembled by
LocalizationFluctuationJointComponents.measurable_toHilbertBlockL2OfComponentsfrom its two vector legs; - each leg is a fixed scalar multiple of a proved measurable class:
Corrector.measurable_freshShellDirichletGradL2for the potential leg, and the differencegrad w_N - (-streamForcing)ofCorrector.measurable_freshShellNeumannGradL2andCorrector.measurable_freshShellForcingL2for the flux leg; - the passage from
cu_Ktoz + cu_nisLocalizationFluctuationJointRestrict.measurable_toHilbertBlockL2OfBlockField_of_subset; - the sample carrier changes from
ShellSeq dtoCutoffSample dalongmeasurable_subtype_coe.
No measurable selection occurs: the corrector families wD, wN are
arbitrary families of weak solutions of e.def.w, exactly as in
Corrector.CorrectorMeasurableGradient, and only their gradient classes are
read.
Main results #
toHilbertBlockL2OfBlockField_congr,toHilbertVectorL2OfVecField_grad,toHilbertVectorL2OfVecField_const_smul--- the class-algebra bookkeeping.memVectorL2_meshCellBackground_potential,..._flux--- the two typing properties, so that a consumer never has to produce them by hand.measurable_toHilbertBlockL2OfBlockField_meshCellBackground--- the class measurability itself, the exact statementLocalizationFluctuationJointBackgroundasks for.measurable_doubledMuValue_meshCellBackground(the measurability half ofhbg) andaemeasurable_doubledMuValue_meshCellBackground_add_testField(hcomp), both in the binder shapeLocalizationFluctuationMuMeasCell.integrable_meshFluctuationCellIntegrand'consumes.
Binders #
Beyond the typing binders: [NeZero d]; the two MeasurableSpace/BorelSpace
pairs on the doubled L^2 carriers (which every consumer in this cone carries,
HilbertBlockL2 having no global Borel instance); the geometric membership of
the cell in the mesh below the localization cube; and the two weak-solution
properties of e.def.w for the corrector families. No integrability, no
estimate, no smallness, no ellipticity and no selection occurs.
Scope #
There is no sorry, no admit, no custom axiom and no set_option maxHeartbeats.
References #
- ABK26,
l.approximate.recurrence.formula, Step 2,e.Pz.def,e.Fz.def, the background display,e.def.w.
Class-algebra bookkeeping #
Two block fields that agree everywhere have the same L^2 class.
The L^2 class of the gradient field of an H^1 function is the proved
gradToHilbertVectorL2.
The class map commutes with a fixed scalar dilation.
A measurable family of classes stays measurable under a fixed scalar dilation.
The two legs of the background, as measurable classes on the large cube #
The potential leg of the background is the corrector gradient itself, so its class is the proved measurable Dirichlet gradient class.
The flux leg of e.Pz.def/e.Fz.def is grad w_N + shom^{-1} h e', whose
class is the difference of the proved Neumann gradient class and the proved
forcing class.
The flux leg's class is measurable in the sample.
The background is cube-average free #
The Step-2 background of the field-slope problem is the affine background.
The localization cube's average enters e.Pz.def with a + and e.Fz.def
with a -, so it cancels in P_z + bfF_z; the resulting field does not depend
on the cell at all. Unconditional.
The potential leg of the background, in closed form.
The flux leg of the background, in closed form.
The typing properties of the background on a cell #
The affine background is vector-L^2 on the localization cube, leg by leg.
Unconditional.
The hpot binder of
LocalizationFluctuationMuMeasCell.integrable_meshFluctuationCellIntegrand', as
a theorem. Conditional only on the geometric membership of the cell.
The hflux binder of
LocalizationFluctuationMuMeasCell.integrable_meshFluctuationCellIntegrand', as
a theorem.
The background's class on the localization cube #
The affine background's L^2 class on the localization cube is a
measurable function of the sample.
The constant bfAhom^{-1/2}(e'; e) is split off by principalLoad_add; what
is left is the doubled class of the two scaled corrector gradients, whose
legs are the proved Corrector.measurable_freshShell{Dirichlet,Neumann}GradL2
and Corrector.measurable_freshShellForcingL2.
only on the two weak-solution properties of e.def.w; the corrector families
are otherwise arbitrary.
The class on a mesoscopic cell #
The remaining obligation of LocalizationFluctuationJointBackground,
discharged.
The L^2 class of the Step-2 background P_z + bfF_z on the mesoscopic cell
z + cu_n is a measurable function of the sample.
No measurability, continuity or selection property of the correctors is assumed, and no minimizer is chosen.
The two consumers #
The measurability half of the hbg binder of
LocalizationFluctuationMuMeasCell.integrable_meshFluctuationCellIntegrand'.
only on the geometric membership of the cell and the two weak-solution
properties of e.def.w.
The hcomp binder of
LocalizationFluctuationMuMeasCell.integrable_meshFluctuationCellIntegrand',
discharged. Every doubled test field on the cell is admissible at once, and no
competitor is selected.