Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section3.Provider.Diffusivity.ApproximateRecurrence.Closure.ClosureInteriorSplit

ClosureInteriorSplit #

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

ClosureSplitInput from an interior-mesh oscillation envelope plus the strip #

ABK26, Step 2 of l.approximate.recurrence.formula, target e.lower.bound.localization.terms.

What this module composes #

Step 2 at the closure's own corrector families runs off a Besov envelope on the full mesh descendantsAtDepth (cu_K) j. The interior route of Closure.GammaTenInterior* and Closure.GammaTenStripEnvelope produces the envelope only on the interior mesh interiorMesoCubeGrid d K (K - j) n --- the carrier forces --- and Closure.GammaTenStripAssembly.fluctuationEnergyAverage_le_of_interior_cell_fold_grid_load is the bridge from the one to the other, at the price of two exposed binders: the strip second moment and its absorption gate.

This module runs Step 2 through that bridge, so the envelope may be read on the interior mesh throughout. The two carrier-facing points are:

The window scale, and the two parameters the bridge needs #

The bridge's scale bookkeeping is K = (K - j) + p, outer + 1 = (K - j) + g, g <= p. At the closure's own data these are forced:

  K - j = recurrenceMesoScale 32 cgamma (n + h) h = n - recurrenceGap 32 cgamma ,
  p = j ,   outer = n ,   g = closureStripGap M = recurrenceGap 32 cgamma + 1 ,

outer = n being exactly the window scale at which exists_besovEnvelope_interiorMesoCubeGrid_of_depth_gridRoot_const produces the interior envelope, one larger than the depth endpoints' own n - 1. g <= p and the two largeness gates hold for every sufficiently large localization cube.

What is proved #

The two named inputs, and their status #

ClosureInteriorBesovEnvelopeInput and ClosureStripRemainder are conditional A obligations of this module; neither is proved here.

Binders #

The two named inputs and the standing 2 <= d. No smallness gate on cgamma, no measurability of the cell integrand and no integrability of the Step-2 integrand is assumed; every threshold is produced.

Scope #

Internal Provider infrastructure. There is no sorry, no admit, no custom axiom and no set_option maxHeartbeats.

References #

The full-mesh to interior-mesh transfer of a grid fourth-moment root #

theorem Algsuperdiff.Section3.Provider.Diffusivity.ApproximateRecurrence.Closure.gridFourthMomentRoot_interiorMesoCubeGrid_le {Omega : Type u_1} [MeasurableSpace Omega] (mu : MeasureTheory.Measure Omega) (d : ℕ) {K nmesh outer : ℤ} (hn : nmesh ≤ K - 1) (houter : outer + 1 ≤ K - 1) (F : HCPolySupport.TriadicCube d → Omega → ℝ) {C : ℝ} (hC : 0 ≤ C) (hroot : gridFourthMomentRoot mu (mesoCubeGrid d K nmesh) F ≤ C) :
gridFourthMomentRoot mu (interiorMesoCubeGrid d K nmesh outer) F ≤ 3 ^ d * C

A grid fourth-moment root survives the passage to the interior mesh, at the absolute factor 3^d.

The interior mesh is a subfamily of the full mesh whose cardinality is at least 3^{-d} of it (LocalizationGrid.card_mesoCubeGrid_le_three_pow_mul_card_interior), so the interior average of the nonnegative cell fourth moments is at most 3^d times the full one. Unconditional beyond the two scale gates and 0 <= C.

The two named inputs of the interior route #

The boundary-strip gap of the closure's own mesh: with the mesh scale K - j = n - recurrenceGap 32 cgamma and window scale outer = n, the bridge's outer + 1 = (K - j) + g forces g = recurrenceGap 32 cgamma + 1.

Equations
  • One or more equations did not get rendered due to their size.
Instances For

    The residual oscillation input of Step 2, on the interior mesh.

    The residual oscillation obligation of Step 2, stated not on the full mesh descendantsAtDepth (cu_K) (closureMeshDepth M n h K) but on the interior mesh interiorMesoCubeGrid d K (K - closureMeshDepth M n h K) n at window scale n, which is the carrier the interior route of Closure.GammaTenInterior* and Closure.GammaTenStripEnvelope produces. Everything else --- the exponent gammaTenBesovExponent = 1/2, the gauge shom_{n+h-1}, the two legs of the gauged bfF_z, the L^4 membership and the cgamma^{15} root --- is unchanged.

    It is a conditional A obligation of this module and is not proved here.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      The boundary-strip residual of the interior route.

      At the closure's own families, for all sufficiently large localization cubes, a bound Cstrip on the full-mesh average of the squared cell expectations of the Step-2 integrand, together with the absorption of the signed Cauchy--Schwarz remainder sqrt(d 3^{g-p}) sqrt(Cstrip) into half of cgamma^{10}. The Filter.atTop quantifier is exactly the one the consumer uses, so nothing is asked for at a cube the consumer does not reach.

      This is exactly what Closure.GammaTenStripAssembly isolates as hstrip and hgate and does not prove; see that module's header. It is a conditional A obligation of this module and is not proved here.

      Equations
      • One or more equations did not get rendered due to their size.
      Instances For

        Step 2 from the two inputs #

        Step 2 at the closure's own corrector families, through the strip bridge, i.e. the verbatim antecedent of Closure.SplitFoldIntegrability.closureSplitInput_of_gammaTenBound.

        exactly on ClosureInteriorBesovEnvelopeInput d, ClosureStripRemainder d and the standing 2 <= d.

        Closure.ClosureFinal.ClosureSplitInput from the interior-mesh envelope and the strip residual.