Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section4.Provider.Annular.SubConstFamily

SubConstFamily #

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

The triadic coefficient family of a_L − C at a constant skew matrix C #

ABK26, Section 4.1 abbreviates,

  ã_{L,m}  :=  a_L − (k_L − k_m)_{□_m}

and then reads the response functional J(z+□_n, …; ã_{L,m}) at every cube z + □_n of the scale-n lattice. The subtracted matrix is a constant: the average is taken over the fixed cube □_m, not over the moving cube z + □_n. Consequently ã_{L,m} is a single coefficient field on all of space, and it carries a genuine Chapter 2 triadic coefficient family — one coefficient object per cube, all with the same representative.

This is what Support.fluxCorrectedCoeffOn is not: there the average is taken over the very cube that carries the object, so the field moves with the cube and no family exists. The two constructions share their ellipticity mechanism, and this module re-derives it at a free constant.

The ellipticity mechanism #

a_L = ν Id + k_L with k_L antisymmetric, so symmPart a_L = ν Id (Cutoff.symmPart_coefficientCutoff). Subtracting an antisymmetric constant C does not move the symmetric part, so symmPart (a_L − C) = ν Id still, and the lower ellipticity constant is still the molecular diffusivity ν. The upper constant comes from the entry envelope Cutoff.coefficientCutoffCubeEntryBound M L ω Q + matAbsSum C, fed to the public certificate Cutoff.isEllipticMatrix_of_symmPart_and_entry_bound.

Both helpers of the two existing instances of this mechanism (Cutoff.coefficientCutoffCoeffOn, Support.fluxCorrectedCoeffOn) are private in their own files, so they are re-derived here rather than reused; the de-privatisation is a queued cleanup item elsewhere and is deliberately not performed by this module, which edits no existing file.

What is proved #

References #

The ℓ¹ entry mass of a constant matrix #

The ℓ¹ entry mass of a matrix. It dominates every entry, and is the only new ingredient the ellipticity of a_L − C needs.

Equations
Instances For

    The field a_L − C #

    The manuscript's ã_{L,m} = a_L − (k_L − k_m)_{□_m}, written at a free constant C: a single coefficient field on all of space.

    Equations
    Instances For

      Subtracting an antisymmetric constant does not move the symmetric part: it is still exactly ν Id. This is the whole content of the ellipticity of ã_{L,m}.

      The ellipticity constants on one cube #

      The uniform entry envelope of a_L − C on Q.

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

        The upper ellipticity constant of a_L − C on Q, produced by the proved skew-plus-entry-bound certificate.

        Equations
        Instances For

          The coefficient object and the triadic family #

          The Chapter 2 coefficient object of ã_{L,m} = a_L − C on the cube Q. The lower ellipticity constant is still the molecular diffusivity ν, because the subtracted matrix is antisymmetric.

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

            The triadic coefficient family of ã_{L,m}. Every cube receives the same literal representative a_L − C; only the locally derived upper ellipticity witness changes with the cube, so compatibility is reflexivity.

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