Dimensional densities #
This file defines dimensional density ratios and their upper and lower limits, then proves their basic order-theoretic properties.
The s-density ratio of an outer measure μ at point x with radius r:
μ(B̄(x, r)) / (2r) ^ s.
The closed ball is taken in the extended-distance sense, Metric.closedEBall x (ofReal r),
so that the definition makes sense in a PseudoEMetricSpace. For 0 ≤ r in a (pseudo)
metric space this is the usual closed ball Metric.closedBall x r, by
Metric.closedEBall_ofReal.
Equations
- dimensionalDensityRatio μ s x r = μ (Metric.closedEBall x (ENNReal.ofReal r)) / ENNReal.ofReal ((2 * r) ^ s)
Instances For
Upper s-density of μ at x:
limsup_{r → 0⁺} μ(B̄(x, r)) / (2r) ^ s.
Equations
- dimensionalUpperDensity μ s x = Filter.limsup (dimensionalDensityRatio μ s x) (nhdsWithin 0 (Set.Ioi 0))
Instances For
Lower s-density of μ at x:
liminf_{r → 0⁺} μ(B̄(x, r)) / (2r) ^ s.
Equations
- dimensionalLowerDensity μ s x = Filter.liminf (dimensionalDensityRatio μ s x) (nhdsWithin 0 (Set.Ioi 0))
Instances For
The s-dimensional density of μ at x exists if the density ratio converges as r → 0⁺.
- exists_tendsto : ∃ (y : ENNReal), Filter.Tendsto (dimensionalDensityRatio μ s x) (nhdsWithin 0 (Set.Ioi 0)) (nhds y)
Instances
The s-dimensional density of μ at x (Definition 1.5).
When the limit exists (HasDensity μ s x), this equals
lim_{r → 0⁺} μ(B̄(x, r)) / (2r) ^ s.
When the limit does not exist, this returns a junk value.
Equations
- dimensionalDensity μ s x = sInf {y : ENNReal | Filter.Tendsto (dimensionalDensityRatio μ s x) (nhdsWithin 0 (Set.Ioi 0)) (nhds y)}
Instances For
Comparison between lower and upper density: Θ_*^s(μ, x) ≤ Θ^{*s}(μ, x).
Non-negativity of lower density: 0 ≤ Θ_*^s(μ, x).
Non-negativity of upper density: 0 ≤ Θ^{*s}(μ, x).
If the lower and upper densities are equal, then the density limit exists.
Uses tendsto_of_liminf_eq_limsup from Mathlib, which directly gives convergence
when liminf = limsup in a conditionally complete linear order with order topology
(which ℝ≥0∞ satisfies).
Lemma 2.4: If Θ_*^s(μ, x) ≥ α, then Θ^{*s}(μ, x) ≥ α.
Lemma 2.5: If Θ^{s}(μ, x) ≤ α, then Θ_^s(μ, x) ≤ α.
Lemma 2.6: If Θ_*^s(μ, x) > α, then eventually (for small enough r > 0) the density ratio is > α.
Lemma 2.7: If Θ^{*s}(μ, x) < α, then eventually (for small enough r > 0) the density ratio is < α.
Lemma 2.8: If Θ_*^s(μ, x) < α, then frequently (along r → 0⁺) the density ratio is < α.
Lemma 2.9: If Θ^{*s}(μ, x) > α, then frequently (along r → 0⁺) the density ratio is > α.
In a (pseudo) metric space, and for a nonnegative radius, the density ratio is computed with the usual closed ball.