Proof of Proposition 2.1, first two equalities (p. 8) — G(t − τᵢ − wᵢ) − G(t − τᵢ) is the F-mass of the delay window
ProvedReedGGN.SystemEq.prop_2_1_arrival_delayLet be a service-time distribution on , with tail , and fix a sample path of the queue with arrival times , nonnegative waiting times and arrival counting process . Then for every ,
Each difference of tails is the probability that a service time falls in the window of length ; this is the first step of the proof of Proposition 2.1, which then turns the window into the event "customer is waiting at time ".
Formalization Note The paper's intermediate form is ambiguous when has an atom at ; the statement uses the paper's next form, with the half-open window, which fixes the convention. is the integral over , the point included. is a probability measure on with and for every real .
import Mathlib import Definitions.Def_ReedGGN_SystemEq_QueueLength
namespace ReedGGN.SystemEq
open MeasureTheory
/-- Proof of Proposition 2.1, first two equalities (p. 8): for each arrival `i ≤ A(t)`, the
difference `G(t − τ_i − w_i) − G(t − τ_i)` is the `F`-mass of the delay window
`{s ≥ 0 : t − (τ_i + w_i) < s ≤ t − τ_i}`, so
`Σ_{i=1}^{A(t)} (G(t − τ_i − w_i) − G(t − τ_i))
= Σ_{i=1}^{A(t)} ∫_0^∞ 1{t − (τ_i + w_i) < s ≤ t − τ_i} dF(s)`. -/
theorem prop_2_1_arrival_delay (μ : Measure ℝ) [IsProbabilityMeasure μ] (hμ : μ (Set.Iio 0) = 0)
(P : SamplePath) (t : ℝ) (ht : 0 ≤ t) :
∑ i ∈ Finset.Icc 1 (P.A t), (G μ (t - P.τ i - P.w i) - G μ (t - P.τ i)) =
∑ i ∈ Finset.Icc 1 (P.A t),
∫ s in Set.Ici 0, (Set.Ioc (t - (P.τ i + P.w i)) (t - P.τ i)).indicator
(fun _ => (1 : ℝ)) s ∂μ := by sorry
end ReedGGN.SystemEq
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.