Eq. (3.9): the geometric steady-state solution of the partial-batch M/M^[K]/1 queue
ProvedQueueingFundamentals.AdvMarkov.partial_batch_steady_stateConsider the partial-batch bulk-service queue with arrival rate , batch service rate and maximal batch size , and assume . Then:
- the characteristic equation of (3.8),
has exactly one root in the open interval ; 2. for this root, the sequence
is a probability distribution solving the balance equations (3.7), and every probability solution of (3.7) equals it.
The steady state thus has the same geometric form as that of the queue, with in place of ; for , .
Formalization Note The book invokes Rouché's theorem to assert that exactly one root lies in without naming the condition; (the server's capacity exceeds the arrival rate) is exactly the condition under which such a root exists, and is stated as a hypothesis. Both the existence and the uniqueness of the root are part of the statement.
import Mathlib import Definitions.Def_QueueingFundamentals_AdvMarkov_BulkService
namespace QueueingFundamentals.AdvMarkov
/-- Eq. (3.9), p.124: for the partial-batch `M/M^[K]/1` queue with `λ < Kμ`, the characteristic
equation `μr^{K+1} − (λ + μ)r + λ = 0` of (3.8) has exactly one root `r_0` in `(0, 1)`, and the
steady-state solution of (3.7) is `p_n = (1 − r_0)r_0ⁿ` (`n ≥ 0`): this sequence is a probability
solution of (3.7), and every probability solution of (3.7) equals it. -/
theorem partial_batch_steady_state (lam mu : ℝ) (K : ℕ) (hlam : 0 < lam) (hmu : 0 < mu)
(hK : 1 ≤ K) (hstab : lam < (K : ℝ) * mu) :
(∃! r0 : ℝ, r0 ∈ Set.Ioo (0 : ℝ) 1 ∧ mu * r0 ^ (K + 1) - (lam + mu) * r0 + lam = 0) ∧
∀ r0 : ℝ, r0 ∈ Set.Ioo (0 : ℝ) 1 → mu * r0 ^ (K + 1) - (lam + mu) * r0 + lam = 0 →
IsPartialBatchSteadyState lam mu K (fun n : ℕ => (1 - r0) * r0 ^ n) ∧
∀ p : ℕ → ℝ, IsPartialBatchSteadyState lam mu K p →
∀ n : ℕ, p n = (1 - r0) * r0 ^ n := by sorry
end QueueingFundamentals.AdvMarkov
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.