§4, p. 135 — uniqueness of the equilibrium distribution (under bounded arrival rates)
ProvedJacksonJobshop.Equilibrium.equilibrium_uniqueLet be a jobshop-like queueing system satisfying Assumptions (2.1)–(2.4), and suppose its arrival rates are bounded: there is with for all . Then the system has at most one equilibrium state probability distribution: if and are probability distributions over state vectors that are both constant solutions of the balance equations (3.1), then
Together with existence, this is what makes "the" equilibrium distribution of Theorem (4.5) well defined.
Formalization Note The paper asserts uniqueness without proof, citing a limit theorem for regular processes; we state uniqueness under bounded arrival rates, which holds for every example in the paper. Bounded arrival rates make the underlying Markov process non-explosive; without such a condition the model admits explosive systems with (e.g. , , ), for which the paper's cited argument does not apply. Equations (3.1) are taken with the arrival-outflow coefficient (see the system definition).
import Mathlib import Definitions.Def_JacksonJobshop_Equilibrium_System
namespace JacksonJobshop.Equilibrium
/-- Jackson (1963), p. 135, §4 opening paragraph: an equilibrium state probability distribution,
if it exists, is unique. Stated under bounded arrival rates `λ(K) ≤ Λ` (an added hypothesis:
the paper asserts uniqueness without proof, citing a limit theorem for regular processes). -/
theorem equilibrium_unique {N : ℕ} (sys : JobshopSystem N)
(hbdd : ∃ Λ : ℝ, ∀ K, sys.lam K ≤ Λ) :
∀ q q' : (Fin N → ℕ) → ℝ, IsEquilibrium sys q → IsEquilibrium sys q' → q = q' := by sorry
end JacksonJobshop.Equilibrium
Read-back
What the Lean code literally says, in plain math · claude-opus-5-5
Setting. Take any and any jobshop system with centers. Its assumptions are:
- ;
- for all , or is positive up to some and zero after it;
- and for ;
- is non-negative, and each row sums to over targets ;
- is the unique solution of .
Hypothesis. There is a real with for all .
Equilibrium. A real function on is an equilibrium if all of the following hold:
- ;
- is summable with sum exactly ;
- at every , the following balance equation holds:
Claim. Any two equilibria are equal at every state. The statement does not assert that an equilibrium exists.
Degenerate cases. If no equilibrium exists, the claim is vacuous. The boundedness hypothesis is automatically satisfied when is zero beyond some , because then takes only finitely many nonzero values. It is a real restriction only in the case where is positive everywhere.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.