Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2.9 — the reversed process is an open migration process

Proved
KellyStochasticNetworks.open_migration_reversed

by naimengye · Sep 18, 2026 · Mathlib 0df444a (Lean v4.33.1)

markov-chainqueueing-theoryreversibilitystochastic-networks

An open migration process is not reversible in general — the detailed balance equations fail — but it behaves well under time reversal. Let π\piπ be its product-form equilibrium distribution and let

q′(n,m)=π(m) q(m,n)π(n)q'(n,m) = \frac{\pi(m)\,q(m,n)}{\pi(n)}q′(n,m)=π(n)π(m)q(m,n)​

be the transition rates of the reversed process, as computed in Proposition 1.1. Then q′q'q′ is again of open-migration form, with the three rate families

  1. q′(n,Tjkn)=λjk′ φj(nj)q'(n, T^{jk}n) = \lambda'_{jk}\,\varphi_j(n_j)q′(n,Tjkn)=λjk′​φj​(nj​) for j≠kj \ne kj=k and nj≥1n_j \ge 1nj​≥1, where λjk′=αkαjλkj\displaystyle \lambda'_{jk} = \frac{\alpha_k}{\alpha_j}\lambda_{kj}λjk′​=αj​αk​​λkj​;
  2. q′(n,Tj→n)=μj′ φj(nj)q'(n, T^{j\to}n) = \mu'_j\,\varphi_j(n_j)q′(n,Tj→n)=μj′​φj​(nj​) for nj≥1n_j \ge 1nj​≥1, where μj′=νjαj\displaystyle \mu'_j = \frac{\nu_j}{\alpha_j}μj′​=αj​νj​​;
  3. q′(n,T→kn)=νk′q'(n, T^{\to k}n) = \nu'_kq′(n,T→kn)=νk′​, where νk′=αkμk\nu'_k = \alpha_k \mu_kνk′​=αk​μk​.

The same functions φj\varphi_jφj​ appear, so the reversed process is an open migration process with the same service mechanism and rerouted traffic. Because the reversed departure stream from colony kkk is the forward immigration stream into it, the third identity is what yields Corollary 2.10: the exit process from colony kkk is Poisson of rate αkμk\alpha_k\mu_kαk​μk​.

Formalization Note The three identities are stated pointwise rather than as an equality of rate matrices, with j≠kj \ne kj=k and nj≥1n_j \ge 1nj​≥1 where the target state must be a genuine state of Z+J\mathbb{Z}_+^JZ+J​. The rate matrix is the assembled one, so each identity also asserts that no other transition of the original process happens to land on the same state with a non-zero rate.

Preamble
import Mathlib
import Definitions.Def_KellyStochasticNetworks_Balance
import Definitions.Def_KellyStochasticNetworks_Migration
Formal statement
namespace KellyStochasticNetworks

theorem open_migration_reversed {J : ℕ} (lam : Fin J → Fin J → ℝ)
    (mu nu α g : Fin J → ℝ) (φ : Fin J → ℕ → ℝ)
    (hlamdiag : ∀ j, lam j j = 0)
    (hφ0 : ∀ j, φ j 0 = 0) (hφpos : ∀ j r, 1 ≤ r → 0 < φ j r)
    (hα : ∀ j, 0 < α j) (hg : ∀ j, 0 < g j) :
    (∀ (n : Fin J → ℕ) (j k : Fin J), j ≠ k → 1 ≤ n j →
        reversedRates (openMigrationPi α g φ) (openMigrationRates lam mu nu φ) n (Tjk j k n)
          = (α k * lam k j / α j) * φ j (n j))
      ∧ (∀ (n : Fin J → ℕ) (j : Fin J), 1 ≤ n j →
        reversedRates (openMigrationPi α g φ) (openMigrationRates lam mu nu φ) n (Tout j n)
          = (nu j / α j) * φ j (n j))
      ∧ (∀ (n : Fin J → ℕ) (k : Fin J),
        reversedRates (openMigrationPi α g φ) (openMigrationRates lam mu nu φ) n (Tin k n)
          = α k * mu k) := by sorry

end KellyStochasticNetworks
Source
Kelly & Yudovina, Stochastic Networks, CUP 2014, pp. 32-33 (PDF pp. 40-41), Theorem 2.9: 'If (n(t), t in R) is a stationary open migration process, then so is the reversed process (n(-t), t in R).' Proof: 'Using Proposition 1.1, we have q'(n, T^{jk}n) = ... = lambda'_{jk} phi_j(n_j), where lambda'_{jk} = (alpha_k/alpha_j) lambda_{kj}. Similarly we find (check this) that the remaining transition rates of the reversed open migration process have the form q'(n, T^{j->}n) = mu'_j phi_j(n_j), q'(n, T^{->k}n) = nu'_k, where mu'_j = nu_j/alpha_j and nu'_k = alpha_k mu_k.' sha256 ec271d555059aee58613e5e9a98b8346214b16185c527314f5d94c1ac8b17b6a
Human review
  • Endorsed by Shuze Chen · Sep 25, 2026

    Confirmed by the moderator at approval.

  • Endorsed by naimengye · Sep 25, 2026

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

  • 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