Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section4.Provider.ExcessDecay.TranslationTransportAssembly

TranslationTransportAssembly #

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

The equation leg of the interior clause, assembled #

Starting from the frozen theorem's own datum

  IsDirichletSolutionOn a_L(·,ω) □_m u h g

and its own interior gate (z+□_{n+2}) ∩ ∂□_m = ∅, it produces

  IsForcedEquation □_{n+2}
    (fluxCorrectedCoeffFamily M L (n+2) □_{n+2} (translateCutoffSample z ω))
    (u(· + z))  (fun x => -g(x + z))

The three steps:

  1. restriction to z + □_{n+2} — EquationRestriction.lean, driven by the anchor's frontier-empty gate;
  2. untranslation to □_{n+2} at translateCutoffSample z ω — TranslationTransport.lean;
  3. the antisymmetric shift a_L ↦ ã_{L,n+2} — AntisymmetricShift{,Cutoff}.lean.

Step 3 is performed after the untranslation, so the subtracted constant is (k_L − k_{n+2})_{□_{n+2}} computed at the translated sample — which, by the exact translation identity of step 2, is the manuscript's (k_L − k_{n+2})_{z+□_{n+2}}. The order is immaterial (the constant is translation-inert), but this order is the one in which every proved cap is stated.

What the forcing becomes, exactly #

The translated frame's forcing is x ↦ -g(x + z); the sign is CoarseGraining's ∇·g convention (Support/Dirichlet.lean) and the translation is the A4 frame move. Both are invisible to the frozen theorem's right-hand side, whose g-legs are seminorms ([g]_{H^s}), even and translation-covariant.

References #

1. The coefficient field of the flux-corrected family #

The representative of Support.fluxCorrectedCoeffFamily on every cube is the literal flux-corrected field.

2. The assembled equation leg #

3. The transported solution, read pointwise #