Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section3.Provider.ErrorComparison.OneCube

OneCube #

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

Provider: the one-cube descendant bound for the homogenization error #

This file proves a local version of the display e.bound.one.cube.by.mathcalE of ABK26: for every k ∈ (-∞,m] ∩ ℤ,

max_{z ∈ 3^k ℤ^d ∩ □_m} 𝓔_{s,∞,q}(z + □_k; a, a₀) ≤ 3^{s(m-k)} 𝓔_{s,∞,q}(□_m; a, a₀), avg_{z ∈ 3^k ℤ^d ∩ □_m} 𝓔_{s,p,q}(z + □_k; a, a₀) ≤ 3^{s(m-k)} 𝓔_{s,p,q}(□_m; a, a₀),

with the ambient quantifiers of the source, i.e. m ∈ ℤ, p, q ∈ [1,∞] and s ∈ (0,∞).

Scope #

The per-descendant form of the first line of the display is proved here at p = ∞ for every finite q ∈ [1,∞); CoarseGraining previously had only the case q = 1 (Homogenization.Book.Ch02.homogenizationErrorOnCube_infinity_one_descendantsAtScale_le).

Main results #

References #

Comparing the scale responses of a cube and of a descendant #

The finite-q endpoint p = ∞ #