Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section5.Provider.HomogenizationCorrectorPointwise

HomogenizationCorrectorPointwise #

Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.

The cut-off observables of the displacement bounds, with their pointwise corrector bounds #

The stopped-moment bounds on a triadic cube consume an observable together with a bound, at every point, on its distance to a bounded Borel function vanishing at the centre. This file produces, for the cut-off linear and quadratic data c · (e ⬝ x) and c · (e ⬝ x)² of a smooth cut-off c equal to one on the cube centred at the origin, the homogenization observable of the cube with all its properties, and the pointwise corrector bounds:

The corrected observable solves the forced Dirichlet problem because the expected exit time is the zero-trace solution of -∇ · a ∇w = 1, and the divergence of (e ⬝ x) e is e ⬝ e.

Main results #

References #

1. The corrected observable solves the quadratic problem #

The divergence of the field x ↦ (e ⬝ x) e is e ⬝ e.

The coordinates of the field x ↦ (e ⬝ x) e are smooth.

The corrected observable solves the quadratic problem. If u = h + w solves the homogeneous equation on the cube with the trace of the quadratic datum, then u − 2 sigmaBar (e ⬝ e) E, with E the zero-trace solution of -∇ · a ∇E = 1, solves the problem with the forcing -2 sigmaBar (e ⬝ x) e.

2. The cut-off linear datum #

theorem Algsuperdiff.Section5.Provider.exists_cutoffAffineObservable_stream {d : ℕ} [NeZero d] (M : Section3.ABKModel d) (omega : Field.FullSample d M.gamma) (m : ℤ) {sigmaBarM EBm : ℝ} (hEB : 0 ≤ EBm) (hren : ∀ (L : ℤ), m ≤ L → ∀ (u v h : HCPolySupport.H1Function (HCPolySupport.openCubeSet (HCPolySupport.originCube d m))) (g : HCPolySupport.Vec d → HCPolySupport.Vec d) (Kg Kh KhInf : ℝ), Section4.Support.IsDirichletSolutionOn (Section3.Cutoff.coefficientCutoff M.nu L ↑omega).toCoeffField (HCPolySupport.originCube d m) u h g → Section4.Support.IsDirichletSolutionOn (fun (x : HCPolySupport.Vec d) => sigmaBarM • 1) (HCPolySupport.originCube d m) v h g → Section4.Support.HolderSeminormBoundOn (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)) (1 / 2) Kg g → Section4.Support.HolderSeminormBoundOn (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)) (1 / 2) Kh h.grad → (∀ x ∈ HCPolySupport.openCubeSet (HCPolySupport.originCube d m), ‖h.grad x‖ ≤ KhInf) → Section4.Support.HasGradientOn (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)) h.toFun h.grad → ∀ᵐ (x : HCPolySupport.Vec d) ∂MeasureTheory.volume.restrict (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)), Real.rpow 3 (-↑m) * |u.toFun x - v.toFun x| ≤ EBm * (sigmaBarM⁻¹ * Real.rpow 3 (↑m / 2) * Kg + (KhInf + Real.rpow 3 (↑m / 2) * Kh))) (e : HCPolySupport.Vec d) {c : HCPolySupport.Vec d → ℝ} (hc : ContDiff ℝ (↑⊤) c) (hc1 : ∀ x ∈ Support.cubeSetAt 0 m, c x = 1) {N : ℝ} (hbdN : ∀ (x : HCPolySupport.Vec d), |c x * affineObservable e x| ≤ N) :

The observable of the cut-off linear datum, with its pointwise corrector bound. For a smooth cut-off c equal to one on the cube centred at the origin and a uniform bound N for c · (e ⬝ x), there is a homogenization observable of the cube with datum c · (e ⬝ x), and it is everywhere within 3^m EB ‖e‖ of the datum, under the corrector clause of the renormalization theorem at the sample for the scale m.

3. The cut-off quadratic datum #

theorem Algsuperdiff.Section5.Provider.exists_cutoffQuadraticObservable_stream {d : ℕ} [NeZero d] {M : Section3.ABKModel d} {omega : Field.FullSample d M.gamma} (R : MarkovProcess.PositiveC0ContractiveResolvent (HCPolySupport.Vec d)) (hreg : R.OnePointRegular) (hcons : R.kernelSemigroup.IsConservative) (hid : (Field.streamWholeSpaceAnalyticData M omega).KernelResolventIdentifiesAnalyticMinimal R) (hT : ∀ (mu : ↑MarkovProcess.Semigroup.PositiveShift) (g : ZeroAtInftyContinuousMap (HCPolySupport.Vec d) ℝ), (R.operator mu) g = (⋯.resolvent.operator mu) g) (m : ℤ) {sigmaBarM EBm : ℝ} (hsigma : 0 < sigmaBarM) (hEB : 0 ≤ EBm) (hren : ∀ (L : ℤ), m ≤ L → ∀ (u v h : HCPolySupport.H1Function (HCPolySupport.openCubeSet (HCPolySupport.originCube d m))) (g : HCPolySupport.Vec d → HCPolySupport.Vec d) (Kg Kh KhInf : ℝ), Section4.Support.IsDirichletSolutionOn (Section3.Cutoff.coefficientCutoff M.nu L ↑omega).toCoeffField (HCPolySupport.originCube d m) u h g → Section4.Support.IsDirichletSolutionOn (fun (x : HCPolySupport.Vec d) => sigmaBarM • 1) (HCPolySupport.originCube d m) v h g → Section4.Support.HolderSeminormBoundOn (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)) (1 / 2) Kg g → Section4.Support.HolderSeminormBoundOn (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)) (1 / 2) Kh h.grad → (∀ x ∈ HCPolySupport.openCubeSet (HCPolySupport.originCube d m), ‖h.grad x‖ ≤ KhInf) → Section4.Support.HasGradientOn (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)) h.toFun h.grad → ∀ᵐ (x : HCPolySupport.Vec d) ∂MeasureTheory.volume.restrict (HCPolySupport.openCubeSet (HCPolySupport.originCube d m)), Real.rpow 3 (-↑m) * |u.toFun x - v.toFun x| ≤ EBm * (sigmaBarM⁻¹ * Real.rpow 3 (↑m / 2) * Kg + (KhInf + Real.rpow 3 (↑m / 2) * Kh))) (e : HCPolySupport.Vec d) {c : HCPolySupport.Vec d → ℝ} (hc : ContDiff ℝ (↑⊤) c) (hc1 : ∀ x ∈ Support.cubeSetAt 0 m, c x = 1) {N : ℝ} (hbdN : ∀ (x : HCPolySupport.Vec d), |c x * quadraticObservable e x| ≤ N) :

The observable of the cut-off quadratic datum, with the pointwise bound on the corrected observable. For a smooth cut-off c equal to one on the cube centred at the origin and a uniform bound N for c · (e ⬝ x)², there is a homogenization observable u of the cube with datum c · (e ⬝ x)², and at every point of the compactified state space the datum is within 5 (3^m)² EB ‖e‖ (∑ |e_i|) of u − 2 sigmaBar (e ⬝ e) times the expected exit time from the cube, under the corrector clause of the renormalization theorem at the sample for the scale m.