Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2.5 — identifying a stationary distribution from the reversed rates

Proved
SerfozoStochasticNetworks.timeReversal_stationary

by naimengye · Sep 19, 2026 · Mathlib 0df444a (Lean v4.33.1)

markov-chainprobabilitystochastic-systems

Let qqq and qˉ\bar qqˉ​ be rate functions on a state space and let π\piπ be strictly positive. Suppose the families y↦q(x,y)y\mapsto q(x,y)y↦q(x,y), y↦qˉ(x,y)y\mapsto\bar q(x,y)y↦qˉ​(x,y) and y↦π(y)q(y,x)y\mapsto\pi(y)q(y,x)y↦π(y)q(y,x) are summable for every xxx, and that

  • qˉ(x,y)=π(x)−1π(y) q(y,x)\bar q(x,y)=\pi(x)^{-1}\pi(y)\,q(y,x)qˉ​(x,y)=π(x)−1π(y)q(y,x) for all x,yx,yx,y, and
  • ∑yq(x,y)=∑yqˉ(x,y)\sum_y q(x,y)=\sum_y\bar q(x,y)∑y​q(x,y)=∑y​qˉ​(x,y) for every xxx — the two rate functions have the same total exit rate from each state.

Then π\piπ is an invariant measure of qqq:

π(x)∑yq(x,y)=∑yπ(y)q(y,x)for every x.\pi(x)\sum_y q(x,y)=\sum_y\pi(y)q(y,x)\qquad\text{for every }x .π(x)y∑​q(x,y)=y∑​π(y)q(y,x)for every x.

This is a way to find an equilibrium measure by guessing the time-reversed process: propose a qˉ\bar qqˉ​ and a π\piπ, check the two displayed conditions, and the balance equations come for free. Reversibility is not involved — qˉ\bar qqˉ​ need not equal qqq.

Formalization Note The conclusion is the balance equations, which is the content of the book's proof. Calling π\piπ "the stationary distribution" additionally requires it to be normalized and the process to be ergodic, neither of which is an algebraic consequence of the hypotheses; and the statement that qˉ\bar qqˉ​ is the rate function of a time reversal requires the process to be constructed, which is not done here.

The summability hypotheses are explicit because the sums are unconditional sums over an arbitrary state space, which are 000 for a non-summable family; with them the rearrangement in the book's one-line proof is valid.

Preamble
import Mathlib
import Definitions.Def_SerfozoStochasticNetworks_Reversible
Formal statement
namespace SerfozoStochasticNetworks

theorem timeReversal_stationary {E : Type*} (q qbar : E → E → ℝ) (π : E → ℝ)
    (hπ : ∀ x, 0 < π x) (hq : ∀ x, Summable (q x)) (hqbar : ∀ x, Summable (qbar x))
    (hin : ∀ x, Summable fun y => π y * q y x)
    (hrev : ∀ x y, qbar x y = (π x)⁻¹ * π y * q y x)
    (hrate : ∀ x, ∑' y, q x y = ∑' y, qbar x y) :
    IsInvariant q π := by sorry

end SerfozoStochasticNetworks
Source
Serfozo, Introduction to Stochastic Networks, Springer 1999, pp. 48-49 (PDF pp. 61-62), Theorem 2.5: "Suppose the Markov process X is ergodic and there exists a positive distribution pi on E and a transition function qbar on E such that qbar(x, y) = pi(x)^{-1} pi(y) q(y, x), x, y in E, (2.7) sum_y q(x, y) = sum_y qbar(x, y), x in E. (2.8) Then pi is the stationary distribution of X. Also, qbar is the transition function of a time reversal of X when X is stationary." Proof: "Under the assumptions, pi(x) sum_y q(x, y) = pi(x) sum_y qbar(x, y) = sum_y pi(y) q(y, x)." sha256 919f20ee082ec19faa80bdd923a5529fce9c4b6d3264fdb77c5efa64256bb463
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).

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