Documentation

BGW.StableTrees.HeightProcessIdentification

Fixed-time height approximation #

Reversal identifies occupation approximations outside the countable set of jumps. Prefix events transport along the equality of restricted-path laws, yielding both almost sure convergence along the public sequence and convergence in measure.

Occupation reversal holds pathwise for paths with nonnegative jumps.

theorem BGW.StableTrees.ae_heightApprox_eq_supOccupation_reversed {θ : ℝ} (hθ : 1 < θ ∧ θ ≤ 2) (t : NNReal) :
∀ᵐ (f : Skorokhod.DLoc) ∂↑(stableLocalLaw θ), ∀ (ε : ℝ), 0 < ε → heightApprox ε f t = supOccupation ε (reversedAt t f) t

Identification (a.s. for each fixed t, all ε > 0 at once): the interval-infimum approximation is the H01 occupation approximation of the reversed process.

The raw filtration is contained in the sigma-algebra of the restricted path.

The occupation at a fixed horizon is measurable from the restricted path.

The supremum local time at a fixed horizon is measurable from the restricted path.

theorem BGW.StableTrees.ae_reversedAt_of_prefix {θ : ℝ} (hθ : 1 < θ ∧ θ ≤ 2) {t : NNReal} (ht : 0 < t) {p : Skorokhod.DLoc → Prop} (hp : MeasurableSet {f : Skorokhod.DLoc | p f}) (ha : ∀ᵐ (f : Skorokhod.DLoc) ∂↑(stableLocalLaw θ), p f) :

A measurable almost sure prefix property transports to the reversed prefix.

The height approximation vanishes at its initial time for every path and cutoff.

theorem BGW.StableTrees.height_zero (θ : ℝ) (f : Skorokhod.DLoc) :
height θ f 0 = 0

The height starts at zero on every path, including exceptional paths.

theorem BGW.StableTrees.ae_tendsto_heightApprox_supEps {θ : ℝ} (hθ : 1 < θ ∧ θ ≤ 2) (t : NNReal) :

A.s. convergence along the deterministic sequence supEps θ, for each fixed t (the null set depends on t).

theorem BGW.StableTrees.heightApprox_tendstoInMeasure {θ : ℝ} (hθ : 1 < θ ∧ θ ≤ 2) (t : NNReal) :
MeasureTheory.TendstoInMeasure (↑(stableLocalLaw θ)) (fun (ε : ℝ) (f : Skorokhod.DLoc) => heightApprox ε f t) (nhdsWithin 0 (Set.Ioi 0)) fun (f : Skorokhod.DLoc) => height θ f t

Convergence in probability for each fixed t along the full filter ε ↓ 0.

theorem BGW.StableTrees.ae_heightRaw_eq {θ : ℝ} (hθ : 1 < θ ∧ θ ≤ 2) (t : NNReal) :

The raw height is a.s. the finite nonnegative height, for each fixed t.