Eq. (2.77) — the transient law of the M/M/∞ queue
ProvedQueueingFundamentals.Transient.mmInf_transientgenerating-functionsmm-infinityp2o-batch-b23bp2o-gran-per-chapterp2o-plan-bookp2o-v1queueingtransient
Let be the arrival rate and the per-customer service rate of an M/M/∞ queue that starts empty, . Put and
Then:
- the solve the forward equations (2.76) on :
- if and otherwise;
- is a probability distribution for every ;
- for and complex with , the generating function is
- every probability solution of (2.76) with the same initial condition coincides with on .
Letting recovers the Poisson steady state (2.57) of the M/M/∞ queue.
Formalization Note The generating-function identity alone would be a Taylor expansion; the statement ties the Poisson law to the system (2.76) it solves, and adds the uniqueness half in the class of probability solutions.
Preamble
import Mathlib import Definitions.Def_QueueingFundamentals_Transient_forwardEquations import Definitions.Def_QueueingFundamentals_Transient_mmInfTransient
Formal statement
namespace QueueingFundamentals.Transient
/-- (2.77) and the M/M/∞ transient law (p.101). For `λ, μ > 0` and `N(0) = 0`:
the functions `p_n(t) = (1/n!) ((1 - e^{-μt}) λ/μ)^n exp(-(1 - e^{-μt}) λ/μ)` solve the forward
equations (2.76) on `[0, ∞)`, satisfy `p_n(0) = [n = 0]`, form a probability distribution for
every `t ≥ 0`, have generating function
`P(z, t) = ∑ p_n(t) zⁿ = exp((z - 1)(1 - e^{-μt}) λ/μ)` for `|z| ≤ 1`, and are the only
probability solution of (2.76) with that initial condition. -/
theorem mmInf_transient (lam mu : ℝ) (hlam : 0 < lam) (hmu : 0 < mu) :
IsForwardSolution (mmInfRHS lam mu) (mmInfTransient lam mu) ∧
(∀ n : ℕ, mmInfTransient lam mu n 0 = if n = 0 then 1 else 0) ∧
IsProbabilityFamily (mmInfTransient lam mu) ∧
(∀ t : ℝ, 0 ≤ t → ∀ z : ℂ, ‖z‖ ≤ 1 →
HasSum (fun n : ℕ => (mmInfTransient lam mu n t : ℂ) * z ^ n)
(Complex.exp ((z - 1) * ((1 - Real.exp (-mu * t)) * (lam / mu) : ℝ)))) ∧
∀ q : ℕ → ℝ → ℝ, IsForwardSolution (mmInfRHS lam mu) q → IsProbabilityFamily q →
(∀ n : ℕ, q n 0 = if n = 0 then 1 else 0) →
∀ n : ℕ, ∀ t : ℝ, 0 ≤ t → q n t = mmInfTransient lam mu n t := by sorry
end QueueingFundamentals.Transient
Source
Gross, Shortle, Thompson & Harris, Fundamentals of Queueing Theory, 4th ed., Wiley 2008, DOI 10.1002/9781118625651, p.101, Eqs. (2.76)–(2.77) and the expansion of (2.77), §2.11.3
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.