Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 1.3.1 — the Swiss army formula of Palm calculus

Disproved
PalmQueueing.Palm.swiss_army_formula

by mikedeng1 · Oct 1, 2026 · Mathlib 0df444a (Lean v4.33.1)

littles-lawpalm-calculuspoint-processqueueingrate-conservationstationary-process

Theorem 1.3.1 (the Swiss army formula). In the setting of §1.3.7 — arrivals {Tn}\{T_n\}{Tn​} with counting measure AAA and intensity λA\lambda_AλA​, departures {τn}\{\tau_n\}{τn​} with counting measure DDD and no ordering assumed, sojourn times Wn=τn−Tn≥0W_n = \tau_n - T_n \ge 0Wn​=τn​−Tn​≥0 (1.3.26) forming a sequence of marks of AAA, the number in system {X(t)}\{X(t)\}{X(t)} with X(b)−X(a)=A((a,b])−D((a,b])X(b) - X(a) = A((a,b]) - D((a,b])X(b)−X(a)=A((a,b])−D((a,b]) (1.3.27), a non-decreasing corlol integrator {B(t)}\{B(t)\}{B(t)} and a non-negative process {Z(t)}\{Z(t)\}{Z(t)} — suppose that {Z(t)},A,D\{Z(t)\}, A, D{Z(t)},A,D and {X(t)}\{X(t)\}{X(t)} are θt\theta_tθt​-compatible, where {θt}\{\theta_t\}{θt​} is a measurable flow on (Ω,F,P)(\Omega, \mathcal{F}, P)(Ω,F,P) and PPP is θt\theta_tθt​-invariant. Then for all ttt,

λA EA0[∫(0,W0]Z(s) dB(s)]  =  1t E[∫(0,t]X(s−) Z(s) dB(s)].(1.3.28)\lambda_A\, E^0_A \Big[ \int_{(0, W_0]} Z(s)\, dB(s) \Big] \;=\; \frac{1}{t}\, E \Big[ \int_{(0,t]} X(s-)\, Z(s)\, dB(s) \Big] . \tag{1.3.28}λA​EA0​[∫(0,W0​]​Z(s)dB(s)]=t1​E[∫(0,t]​X(s−)Z(s)dB(s)].(1.3.28)

The book's name for it says what it is. Depending on which blade is selected, this one identity gives the main formulas of Palm calculus: with Z≡1Z \equiv 1Z≡1 and B(t)=tB(t) = tB(t)=t it reads λAEA0[W0]=E[X(0)]\lambda_A E^0_A[W_0] = E[X(0)]λA​EA0​[W0​]=E[X(0)], which is Little's law; other choices give the inversion formula, the Miyazawa conservation principle and the rate conservation law.

Two details of the statement are load-bearing. The integrator on both sides is dBdBdB — not dAdAdA, not dDdDdD; {B(t)}\{B(t)\}{B(t)} is introduced two paragraphs above the theorem and shares its symbol with nothing else. And X(s−)X(s-)X(s−) is the left limit, not X(s)X(s)X(s): the proof runs through the Stieltjes-Lebesgue integration-by-parts formula

(f(t)−f(0))(g(t)−g(0))=∫(0,t](f(s−)−f(0)) dg(s)+∫(0,t](g(s)−g(0)) df(s),(1.3.29)(f(t) - f(0))(g(t) - g(0)) = \int_{(0,t]} (f(s-) - f(0))\, dg(s) + \int_{(0,t]} (g(s) - g(0))\, df(s) , \tag{1.3.29}(f(t)−f(0))(g(t)−g(0))=∫(0,t]​(f(s−)−f(0))dg(s)+∫(0,t]​(g(s)−g(0))df(s),(1.3.29)

which needs the predictable version of the integrand.

The compatibility of {B(t)}\{B(t)\}{B(t)} with the flow is assumed here although p.29 does not list it among the compatibility hypotheses; the proof uses it at the step "R(t)=R(0)∘θtR(t) = R(0) \circ \theta_tR(t)=R(0)∘θt​" (p.30), and without it the identity is false, its left-hand side being independent of ttt while its right-hand side would not be. The book states the conclusion "for all t∈Rt \in \mathbb{R}t∈R"; it is stated here for t>0t > 0t>0, the factor 1/t1/t1/t requiring t≠0t \ne 0t=0.

Preamble
import Mathlib
import Definitions.Def_PalmQueueing_Palm_PointProcess
import Definitions.Def_PalmQueueing_Palm_SwissArmySetting

/-!
# Theorem 1.3.1: the Swiss army formula of Palm calculus (§1.3.7, p.29)
-/
Formal statement
namespace PalmQueueing.Palm

open MeasureTheory

variable {Ω : Type*} [MeasurableSpace Ω]

/-- **The Swiss army formula**, Theorem 1.3.1, Eq. (1.3.28) (§1.3.7, p.29). In the setting of
pp.28-29 — arrivals `{T_n}` with counting measure `A` and intensity `λ_A`, departures `{τ_n}` with
counting measure `D` and no ordering assumed, sojourn times `W_n = τ_n - T_n ≥ 0` forming a
sequence of marks of `A`, the number in system `{X(t)}` with `X(b) - X(a) = A((a,b]) - D((a,b])`, a
non-decreasing corlol integrator `{B(t)}` and a non-negative process `{Z(t)}` — for all `t`,

`λ_A E⁰_A [ ∫_{(0, W₀]} Z(s) dB(s) ]  =  (1/t) E [ ∫_{(0, t]} X(s-) Z(s) dB(s) ]`.

The book's name for it says what it is: depending on which blade is selected, this one identity
gives Little's law (`Z ≡ 1`, `B(t) = t`, whereupon it reads `λ_A E⁰_A[W₀] = E[X(0)]`), the
inversion formula, the Miyazawa conservation principle and the rate conservation law.

Two details of the statement are load-bearing. The integrator on **both** sides is `dB`, not `dA`
and not `dD`; and `X(s-)` is the left limit, not `X(s)` — the proof runs through the
Stieltjes-Lebesgue integration by parts formula (1.3.29), which needs the predictable version. -/
theorem swiss_army_formula (S : SwissArmySetting Ω) (t : ℝ) (ht : 0 < t) :
    ENNReal.ofReal S.toPalmSetting.lam *
        ∫⁻ ω, ∫⁻ s in Set.Ioc (0 : ℝ) (S.W 0 ω), ENNReal.ofReal (S.Z s ω)
          ∂((S.B.B ω).measure) ∂S.toPalmSetting.P0
      = (∫⁻ ω, ∫⁻ s in Set.Ioc (0 : ℝ) t, ENNReal.ofReal (S.Xl s ω * S.Z s ω)
          ∂((S.B.B ω).measure) ∂S.toPalmSetting.P) / ENNReal.ofReal t := by sorry

end PalmQueueing.Palm
Source
Baccelli & Bremaud, Elements of Queueing Theory: Palm Martingale Calculus and Stochastic Recurrences, 2nd ed., Springer 2003, p. 29, Theorem 1.3.1, Eq. (1.3.28)
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me