Proof of Theorem 5, pp. 800–801 — the discrete law F₀ satisfies (12b) and (12a)
ProvedBalkemaDeHaan.DiscreteDomain.levels_12b_12aLet , and . Let be a sequence with , and let be a discrete distribution function with jumps at an unbounded increasing sequence , tail equivalent to , with
Then satisfies (12b) and its jumps satisfy (12a):
Combined with the previous milestone this is the "only if" half of Theorem 5.
Formalization Note The page obtains (12a) with limit , the scale constant of the functional equation for the tail of a translate of ; this is the ratio of successive gaps between the jumps of , which is , and the statement says so. The level identity is a hypothesis, as on the page; it cannot be dropped: a law tail equivalent to may carry extra atoms of vanishing relative mass, at which (12a) and (12b) fail.
import Mathlib import Definitions.Def_BalkemaDeHaan_DiscreteDomain_ResidualLife import Definitions.Def_BalkemaDeHaan_DiscreteDomain_PiPC import Definitions.Def_BalkemaDeHaan_DiscreteDomain_DiscreteLaw open MeasureTheory Filter Topology
namespace BalkemaDeHaan.DiscreteDomain
/-- Proof of Theorem 5, pp. 800–801 (PDF 9–10): if `F ∈ D_r(Π_{p,c})` and `F₀` is a discrete law
with jumps `t₀ < t₁ < ⋯ → ∞`, BalkemaDeHaan.LimitTypes.tail equivalent to `F`, with `R₀(t_n) = R(b_n)` for a sequence
with `R(b_{n+1})/R(b_n) → e^{-p}`, then `F₀` satisfies (12b) and its jumps satisfy (12a). -/
theorem levels_12b_12a (p c : ℝ) (hp : 0 < p) (hc : 0 ≤ c)
(μ : Measure ℝ) [IsProbabilityMeasure μ] (hμ : InDr μ (piPC p c))
(b : ℕ → ℝ)
(hb : Tendsto (fun n : ℕ => (μ (Set.Ioi (b (n + 1)))).toReal / (μ (Set.Ioi (b n))).toReal)
atTop (𝓝 (Real.exp (-p))))
(ν₀ : Measure ℝ) [IsProbabilityMeasure ν₀] (t : ℕ → ℝ)
(hν₀ : IsDiscreteWithJumps ν₀ t) (hlev : ∀ n, ν₀ (Set.Ioi (t n)) = μ (Set.Ioi (b n)))
(heq : TailEquiv μ ν₀) :
TailRatio ν₀ t p ∧ GapRatio t p c := by sorry
end BalkemaDeHaan.DiscreteDomain
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.