Lemma 9.3 — if ∫₀^t u^κ P(R > u) du ~ C₊t, then P(R > t) ~ C₊t^{−κ}
ProvedGoldieRenewal.Implicit.lemma_9_3heavy-tailsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1tauberian-theorem
Let be a real random variable, and . If
then
This Tauberian step turns the averaged statement delivered by the key renewal theorem into the pointwise tail asymptotics (2.10); monotonicity of is what makes it possible.
Formalization Note "" is read as and "" as ; for these are and , the paper's reading (p. 130).
Preamble
import Mathlib open MeasureTheory Filter Topology
Formal statement
namespace GoldieRenewal.Implicit
/-- **Lemma 9.3** (Goldie 1991, Ann. Appl. Probab. 1(1), p. 143). If
`∫₀^t u^κ P(R > u) du ~ C₊ t` as `t → ∞`, then `P(R > t) ~ C₊ t^{−κ}` as `t → ∞`.
**Formalization Note** `R` is any real random variable and `κ > 0` (the standing exponent of §2).
"`~ C₊ t`" is read as `(∫₀^t u^κ P(R > u) du) / t → C₊` and "`P(R > t) ~ C₊ t^{−κ}`" as
`t^κ P(R > t) → C₊`, which includes the case `C₊ = 0` read as `o(t)` and `o(t^{−κ})` (the paper's
convention, p. 130). The integrand is bounded and measurable on `[0, t]`, so the interval integral is
the paper's value. -/
theorem lemma_9_3 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(R : Ω → ℝ) (hR : Measurable R) (κ : ℝ) (hκ : 0 < κ) (C : ℝ)
(h : Tendsto (fun t : ℝ => (∫ u in (0 : ℝ)..t, u ^ κ * P.real {ω | u < R ω}) / t)
atTop (𝓝 C)) :
Tendsto (fun t : ℝ => t ^ κ * P.real {ω | t < R ω}) atTop (𝓝 C) := by sorry
end GoldieRenewal.Implicit
Source
Goldie, Implicit renewal theory and tails of solutions of random equations, Ann. Appl. Probab. 1(1):126–166 (1991), DOI 10.1214/aoap/1177005985, p. 143, Lemma 9.3
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.