Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section3.Provider.Scales.Landmarks

Landmarks #

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

Landmark characterizations for the Section 3 induction scales #

ABK26 introduces two distinguished small scales. The scale m** is the largest integer m with nu ^ 2 ≥ 2 (log 3) c gamma⁻³ 3 ^ (2 gamma m) and the scale m* is the largest integer m with nu ^ 2 ≥ 2 (log 3) c gamma⁻¹ 3 ^ (2 gamma m).

This file proves exactly those characterizations for the repository's Algsuperdiff.Section3.mStarStar and Algsuperdiff.Section3.mStar, together with the floor estimate that they imply and the ordering mStarStar ≤ mStar.

The Lean definitions use the normalized Frobenius origin-mass constant Algsuperdiff.Section3.Disorder.cstarPlus where the paper prints cstar. That substitution is already part of the frozen definitions and is not revisited here.

Main results #

References #

Quotient form of the defining property of mStarStar.

Quotient form of the defining property of mStar.

mStar M is the largest integer m with 2 (log 3) c⁺ gamma⁻¹ 3 ^ (2 gamma m) ≤ nu ^ 2, which is ABK26's defining description of m*.

The base-case landmark never exceeds the base-plateau landmark, because gamma ≤ 1 / 4 forces gamma ^ 3 ≤ gamma.

The upper half of the floor sandwich at mStarStar.