Eq. (1.2.25) — the inversion formula of Ryll-Nardzewski and Slivnyak
ProvedPalmQueueing.Palm.inversion_formulapalm-calculuspoint-processqueueingstationary-process
The inversion formula of Ryll-Nardzewski and Slivnyak. The original (-invariant) probability can be recovered from the Palm probability by averaging over one inter-point interval: for all non-negative random variables ,
The book derives it from Mecke's formula applied to for any with -a.s. (1.2.23), specialized to and using that -a.s. for all .
Two specializations appear on the same page: gives (1.2.26), and gives
It is stated for all non-negative random variables, not only bounded ones; restricting it to bounded would be a strictly weaker statement than the book's.
Preamble
import Mathlib import Definitions.Def_PalmQueueing_Palm_PointProcess /-! # Eq. (1.2.25): the inversion formula of Ryll-Nardzewski and Slivnyak (p.20) -/
Formal statement
namespace PalmQueueing.Palm
open MeasureTheory
variable {Ω : Type*} [MeasurableSpace Ω]
/-- **The inversion formula of Ryll-Nardzewski and Slivnyak**, Eq. (1.2.25) (p.20): the
`θ_t`-invariant probability is recovered from the Palm probability by averaging over one inter-point
interval,
`E[f] = λ E⁰_N [ ∫_0^{T₁} (f ∘ θ_t) dt ]`,
for **all** non-negative random variables `f` — not only bounded ones. Taking `f = 1` gives
`λ E⁰_N[T₁] = 1` (1.2.27), and `f = 1_A` gives
`P(A) = λ ∫_0^∞ P⁰_N(T₁ > t, θ_t ∈ A) dt` (1.2.26). -/
theorem inversion_formula (S : PalmSetting Ω) (f : Ω → ENNReal) (hf : Measurable f) :
∫⁻ ω, f ω ∂S.P
= ENNReal.ofReal S.lam *
∫⁻ ω, ∫⁻ t in Set.Ioc (0 : ℝ) (S.N.T 1 ω), f (S.θ t ω) ∂(volume : Measure ℝ) ∂S.P0 := by sorry
end PalmQueueing.Palm
Source
Baccelli & Bremaud, Elements of Queueing Theory: Palm Martingale Calculus and Stochastic Recurrences, 2nd ed., Springer 2003, p. 20, Eq. (1.2.25)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.