Sets with bounded covering exponent #
A property of sets in a pseudoemetric space that bounds the growth of their covering numbers as the radius of the balls goes to zero. This will be used in the proof of the Kolmogorov-Chentsov theorem, that gives a sufficient condition for a stochastic process to have a modification with Hölder continuous paths.
Main definitions #
HasCoveringExponent A c d: a setAin a pseudoemetric space has bounded covering number with constantcand exponentdif it has finite diameter and for allε ∈ (0, diam(A)], the covering number ofAat scaleεis bounded byc * ε^{-d}.
Main statements #
HasCoveringExponent.subset: ifAhas bounded covering exponent with constantcand exponentd, then any subset ofAhas bounded covering exponent with constant2^d * cand exponentd.
structure
Metric.HasCoveringExponent
{T : Type u_1}
[PseudoEMetricSpace T]
(A : Set T)
(c : ENNReal)
(d : ℝ)
:
A set A in a pseudoemetric space has bounded covering number with constant c and exponent
d if it has finite diameter and for all ε ∈ (0, diam(A)], the covering number of A
at scale ε is bounded by c * ε^{-d}.
Instances For
theorem
Metric.HasCoveringExponent.coveringNumber_lt_top
{T : Type u_1}
[PseudoEMetricSpace T]
{A : Set T}
{c : ENNReal}
{ε : NNReal}
{d : ℝ}
(h : HasCoveringExponent A c d)
(hε_ne : ε ≠ 0)
(hc : c ≠ ⊤)
(hd : 0 ≤ d)
:
theorem
Metric.HasCoveringExponent.coveringNumber_ne_top
{T : Type u_1}
[PseudoEMetricSpace T]
{A : Set T}
{c : ENNReal}
{ε : NNReal}
{d : ℝ}
(h : HasCoveringExponent A c d)
(hε_ne : ε ≠ 0)
(hc : c ≠ ⊤)
(hd : 0 ≤ d)
:
theorem
Metric.HasCoveringExponent.subset
{T : Type u_1}
[PseudoEMetricSpace T]
{A B : Set T}
{c : ENNReal}
{d : ℝ}
(h : HasCoveringExponent A c d)
(hBA : B ⊆ A)
(hd : 0 ≤ d)
:
HasCoveringExponent B (2 ^ d * c) d