Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section4.Provider.ExcessDecay.InteriorGlueFrame

InteriorGlueFrame #

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

The off-grid error is a cube error in the translated frame #

It is not new: CoarseGraining already proves it, unconditionally, sorry-free, with oleans built — Homogenization/CoarseGraining/Translation.lean carries the full covariance of the coarse-graining apparatus under a real translation:

  ResponseJ (w + U) p q a = ResponseJ U p q (a(· + w))        (:372)
  BlockJ    (w + U) P Q a = BlockJ    U P Q (a(· + w))        (:303)
  Mu        (w + U) P   a = Mu        U P   (a(· + w))        (:116)
  σ⋆, σ⋆⁻¹, κ, σ, b coarse: the same, at :378--:439 .
  𝓔_{t,∞,2}(w + P; g, a₀) = 𝓔_{t,∞,2}(P; g(·+w), a₀) .

What this unblocks (and what it does not) #

With this bridge, the child-frame composition can feed InteriorGlueCap's cap at the root, provided the child-frame family is built from the parent's flux-corrected field rather than the child's own (the child-frame family uses the child's own antisymmetric increment, which is a different — equally valid — shift of the same equation). That re-basing, the analogous transport of the coarse-grained ellipticity factors λ_{r,q}, Λ_{r,q}, and the energy-average transport are the remaining assembly steps; none of them is assumed here.

References #

1. The adjoint commutes with the translation #

2. The set-level doubled response is translation covariant #

The doubled response of a translated set. Both halves move by CoarseGraining's ResponseJ_translateSet_eq_translateCoeffField; the adjoint half needs the commutation of §1.

3. The one-cube quantity #

4. The shells and the multiscale error #