SmallGammaRpow #
Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.
Smallness of the amplitude γ^{1/2}|log γ|^q #
The accuracy windows of the percolation estimates have a lower endpoint of the form
C γ^{1/2} |log γ|^q
with a real exponent q, and are useful only when that endpoint lies below the fixed
upper endpoints the estimates impose. This file supplies the elementary fact that makes
those windows nonempty: for a fixed multiplier and a fixed exponent the amplitude falls
below any prescribed positive bound once the disorder parameter is small enough.
The proof is quantitative. For 0 < r and 0 < γ ≤ 1 the logarithm obeys
|log γ| ≤ r⁻¹ γ^{-r} ,
which is log x ≤ x - 1 read at x = γ^{-r}. Taking the q-th power with
r = (4(q+1))⁻¹ and multiplying by γ^{1/2} gives
γ^{1/2} |log γ|^q ≤ (4(q+1))^q γ^{1/4} ,
so the threshold can be taken to be a fourth power.
Main results #
abs_log_le_inv_mul_rpow_neg— the logarithm against a small negative power.sqrt_mul_abs_log_rpow_le_rpow_quarter— the amplitude againstγ^{1/4}.exists_sqrt_mul_abs_log_rpow_le— any multiple of the amplitude is below any prescribed positive bound onceγis below an explicit threshold.
The logarithm is dominated by any small negative power. For 0 < γ ≤ 1 and
0 < r,
|log γ| ≤ r⁻¹ γ^{-r} .
The amplitude γ^{1/2}|log γ|^q is a bounded multiple of γ^{1/4}. For every
real exponent q ≥ 0 and every 0 < γ ≤ 1,
γ^{1/2} |log γ|^q ≤ (4(q+1))^q γ^{1/4} .
Any multiple of the amplitude γ^{1/2}|log γ|^q is uniformly small for small
disorder. For every nonnegative multiplier K, every real exponent q ≥ 0 and every
prescribed positive bound eps there is a threshold below which
K γ^{1/2}|log γ|^q ≤ eps .
The threshold is explicit: with C = (4(q+1))^q and B = K C + 1 it is
min 1 ((eps / B)^4), because γ^{1/2}|log γ|^q ≤ C γ^{1/4}.