Finite-first-moment compound-Poisson sample-path ratio converges almost surely
ProvedKyprianou.Ch01.compoundPoissonFiniteOffspringMeanRatioConvergesAEcompound-poissonfinite-first-momentkyprianou-ch01strong-lawtime-asymptotic
For a genuine rate-positive compound Poisson driver whose source offspring law has finite first moment and exact integer-jump support and masses, the real aggregate-jump path divided by one plus physical time has an almost-sure finite real limit. This is the full strong-law component needed for the source-faithful eventual linear upper bound; no limit value is assumed.
Preamble
import Mathlib import Definitions.Def_LeanLevy_Processes_CompoundPoisson open MeasureTheory ProbabilityTheory Set Filter open scoped NNReal ENNReal set_option autoImplicit false
Formal statement
theorem Kyprianou.Ch01.compoundPoissonFiniteOffspringMeanRatioConvergesAE
{Ω : Type*} [MeasurableSpace Ω]
(Q : Measure Ω) [IsProbabilityMeasure Q]
(rate : ℝ≥0) (hrate : 0 < rate)
(offspring : PMF ℕ)
(hFiniteMean : Summable (fun k : ℕ => (offspring k).toReal * (k : ℝ)))
(arrival marks : ℕ → Ω → ℝ)
(jumpLaw : Measure ℝ) [IsProbabilityMeasure jumpLaw]
(hdriver : IsCompoundPoissonDriver arrival marks rate jumpLaw Q)
(hSupport : jumpLaw {x : ℝ | ∃ k : ℕ, x = (k : ℝ) - 1} = 1)
(hJumpMass : ∀ k : ℕ,
jumpLaw ({((k : ℝ) - 1)} : Set ℝ) = offspring k) :
∀ᵐ ω ∂Q, ∃ L : ℝ,
Tendsto (fun t : ℝ≥0 =>
ProbabilityTheory.compoundPoisson 0 arrival marks (t : ℝ) ω /
(1 + (t : ℝ))) atTop (nhds L) := by
sorrySource
Kyprianou (2006), Chapter 1 Section 1.3.4, compound Poisson strong law under finite mean; reusable Markov branching non-explosion dependency.