Documentation

LeanPool.AnomalousDiffusion.AVenhance.Infra.Section5.HMinusTools

Tools for the extended-valued homogeneous negative norm. The mean-zero bridge retains the distinction between all periodic tests and the centred tests used in Ergodic.

The App C analytic, centred test norm is bounded by the smooth-test norm. The reverse inequality would need a density argument; it is not assumed.

Poincare duality in the carriers, with its sharp torus constant.

theorem AVenhance.Infra.Section5.hMinusOneNorm_le_of_divergence {h : Homogenization.Vec 2 → ℝ} {V : Fin 2 → Homogenization.Vec 2 → ℝ} (hVdiff : ∀ (i : Fin 2), ContDiff ℝ (↑⊤) (V i)) (hVper : ∀ (i : Fin 2), Ergodic.IsZPeriodic (V i)) (hdiv : ∀ (x : Homogenization.Vec 2), h x = ∑ i : Fin 2, (fderiv ℝ (V i) x) (Homogenization.basisVec i)) :

Divergence duality for the C-infinity test class. This transfers the integration-by-parts proof underlying the analytic-test estimate, rather than asserting an unproved equality of the two test suprema.

Classical divergence written coordinatewise on the Vec carrier.

Equations
Instances For

    Vector-field form of divergence duality, with the Euclidean energy.

    Triangle inequality; L2 hypotheses justify addition of the dual integrals. Means need not vanish: the extended norm retains infinite values.

    Exact scaling, valid even for infinite negative norm.

    Exactly the time-norm carrier occurring in the BigBound statement.

    Equations
    Instances For

      Time-integrated Minkowski, with explicit measurability of the two norms.

      theorem AVenhance.Infra.Section5.timeHMinusOneNorm_le_of_bound {f : ℝ → Homogenization.Vec 2 → ℝ} {B : ℝ → ENNReal} (h : ∀ t ∈ Set.Ioo 0 1, hMinusOneNorm (f t) ≤ B t) :
      timeHMinusOneNorm f ≤ (∫⁻ (t : ℝ) in Set.Ioo 0 1, B t ^ 2) ^ (1 / 2)

      Pointwise bounds pass to the exact BigBound time norm.

      Time-integrated L2-to-negative-norm estimate in carriers.

      noncomputable def AVenhance.Infra.Section5.periodicPrimitive (g : ℝ → ℝ) (t : ℝ) :

      The time primitive based at 0, without a mean correction.

      Equations
      Instances For

        The primitive is an actual antiderivative, not only a periodic integral.

        A mean-zero period integral makes the primitive periodic.

        theorem AVenhance.Infra.Section5.periodicPrimitive_abs_le {g : ℝ → ℝ} {T B : ℝ} (hT : 0 < T) (hg : Continuous g) (hp : Function.Periodic g T) (hm : ∫ (s : ℝ) in 0..T, g s = 0) (hB : 0 ≤ B) (hb : ∀ t ∈ Set.Icc 0 T, |g t| ≤ B) (t : ℝ) :

        Global supremum bound in a pointwise form, retaining all real times. A bound just on one period is sufficient by periodicity.