Documentation

BGW.MetricMeasure.MeasuredScaling

Positive spatial scaling of unrooted measured GH classes #

The same representative scaling used for rooted compact probability spaces descends to the unrooted whole-carrier measured-isometry quotient. Measure transport is unchanged. As in the rooted case, GHP estimates use max factor 1 because the Lévy–Prokhorov component is not homogeneous under spatial scaling.

theorem BGW.MetricMeasure.CompactMeasuredSpace.measuredIsometry_scale (factor : ℝ) (hfactor : 0 < factor) {X Y : CompactMeasuredSpace} (h : X.MeasuredIsometry Y) :
(scale factor hfactor X).MeasuredIsometry (scale factor hfactor Y)

Whole-carrier measured isometries remain so after positive spatial scaling.

theorem BGW.MetricMeasure.CompactMeasuredSpace.distance_scale_le (factor : ℝ) (hfactor : 0 < factor) (X Y : CompactMeasuredSpace) :
(scale factor hfactor X).distance (scale factor hfactor Y) ≤ max factor 1 * X.distance Y

Positive scaling is Lipschitz for the represented unrooted GHP distance.

Scaling by one preserves the represented unrooted measured isometry class.

theorem BGW.MetricMeasure.CompactMeasuredSpace.measuredIsometry_scale_inv (factor : ℝ) (hfactor : 0 < factor) (X : CompactMeasuredSpace) :
(scale factor⁻¹ ⋯ (scale factor hfactor X)).MeasuredIsometry X

Scaling by the positive reciprocal cancels the first scaling, including the measure.

theorem BGW.MetricMeasure.CompactMeasuredSpace.min_mul_distance_le_scale (factor : ℝ) (hfactor : 0 < factor) (X Y : CompactMeasuredSpace) :
min factor 1 * X.distance Y ≤ (scale factor hfactor X).distance (scale factor hfactor Y)

The inverse scaling gives the lower bound with factor min factor 1.

Changing the positive scaling factor is controlled by the representative diameter.

noncomputable def BGW.MetricMeasure.MeasuredGHSpace.scale (factor : ℝ) (hfactor : 0 < factor) :

Positive spatial scaling on whole-carrier unrooted measured isometry classes.

Equations
Instances For
    @[simp]

    Scaling a quotient class is scaling any of its compact measured representatives.

    @[simp]

    Scaling an unrooted measured GH class by one is the identity.

    @[simp]
    theorem BGW.MetricMeasure.MeasuredGHSpace.scale_inv (factor : ℝ) (hfactor : 0 < factor) (X : MeasuredGHSpace) :
    scale factor⁻¹ ⋯ (scale factor hfactor X) = X

    Scaling by a positive factor and its reciprocal is the identity on unrooted measured classes.

    theorem BGW.MetricMeasure.MeasuredGHSpace.dist_scale_le (factor : ℝ) (hfactor : 0 < factor) (X Y : MeasuredGHSpace) :
    dist (scale factor hfactor X) (scale factor hfactor Y) ≤ max factor 1 * dist X Y

    Positive spatial scaling has the same upper GHP Lipschitz bound as in the rooted case.

    theorem BGW.MetricMeasure.MeasuredGHSpace.min_mul_dist_le_scale (factor : ℝ) (hfactor : 0 < factor) (X Y : MeasuredGHSpace) :
    min factor 1 * dist X Y ≤ dist (scale factor hfactor X) (scale factor hfactor Y)

    Positive spatial scaling has the lower GHP expansion bound min factor 1.

    theorem BGW.MetricMeasure.MeasuredGHSpace.lipschitzWith_scale (factor : ℝ) (hfactor : 0 < factor) :
    LipschitzWith ⟨max factor 1, ⋯⟩ (scale factor hfactor)

    Fixed-factor spatial scaling is Lipschitz for the unrooted measured GHP metric.

    theorem BGW.MetricMeasure.MeasuredGHSpace.continuous_scale (factor : ℝ) (hfactor : 0 < factor) :
    Continuous (scale factor hfactor)

    Fixed-factor spatial scaling is continuous on unrooted measured GH classes.

    theorem BGW.MetricMeasure.MeasuredGHSpace.measurable_scale (factor : ℝ) (hfactor : 0 < factor) :
    Measurable (scale factor hfactor)

    Fixed-factor spatial scaling is Borel measurable on unrooted measured GH classes.

    Spatial scaling is jointly continuous in its positive factor and unrooted measured class.

    Joint spatial scaling is Borel measurable for the product Borel structure.