Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Equation (3.3) — the exact equilibrium distribution of a loss network

Proved
KellyStochasticNetworks.loss_network_equilibrium

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

loss-networksmarkov-chainqueueing-theorystochastic-networksteletraffic

A loss network with fixed routing has JJJ links, link jjj carrying CjC_jCj​ circuits, and RRR routes; a call on route rrr requires AjrA_{jr}Ajr​ circuits from link jjj and is lost if any link has fewer free. Calls on route rrr arrive as a Poisson process of rate νr>0\nu_r > 0νr​>0 and hold their circuits for an exponentially distributed time of unit mean. The feasible states are S(C)={n∈Z+R:An≤C}S(C) = \{n \in \mathbb{Z}_+^R : An \le C\}S(C)={n∈Z+R​:An≤C}.

The process is the uncapacitated network — independent Poisson streams — truncated to S(C)S(C)S(C), so by Lemma 3.4 its equilibrium distribution is

π(n)=G(C)∏rνr nrnr!,n∈S(C),G(C)=(∑n∈S(C)∏rνr nrnr!)−1,\pi(n) = G(C)\prod_{r}\frac{\nu_r^{\,n_r}}{n_r!}, \qquad n \in S(C), \qquad G(C) = \Bigl(\sum_{n \in S(C)}\prod_r \frac{\nu_r^{\,n_r}}{n_r!}\Bigr)^{-1},π(n)=G(C)r∏​nr​!νrnr​​​,n∈S(C),G(C)=(n∈S(C)∑​r∏​nr​!νrnr​​​)−1,

equation (3.3). Precisely: π\piπ satisfies the detailed balance equations for the truncated rates, and sums to 111 over S(C)S(C)S(C).

In words, the equilibrium law of a loss network is that of independent Poisson random variables conditioned on the linear inequalities An≤CAn \le CAn≤C. This is exact, not an approximation. It is also not directly useful for a large or complex network: computing G(C)G(C)G(C) is a sum over the feasible states, and for a general non-negative integer matrix AAA it is NP-hard, which is why the rest of the chapter develops the Erlang fixed point approximation.

Formalization Note The truncated process lives on the subtype of feasible states, so a transition that would leave S(C)S(C)S(C) has no target and is suppressed automatically. Following the book, G(C)G(C)G(C) is the reciprocal of the sum of the weights, so the summability hypothesis is stated for G−1G^{-1}G−1; it asserts convergence and the value at once.

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

theorem loss_network_equilibrium {J R : ℕ} (A : Fin J → Fin R → ℕ) (C : Fin J → ℕ)
    (ν : Fin R → ℝ) (hν : ∀ r, 0 < ν r) (G : ℝ) (hG0 : G ≠ 0)
    (hG : HasSum (fun n : lossStates A C => lossWeight ν (n : Fin R → ℕ)) G⁻¹) :
    DetailedBalance (fun n : lossStates A C => G * lossWeight ν (n : Fin R → ℕ))
        (truncatedRates
          (openMigrationRates (fun _ _ => (0 : ℝ)) (fun _ => 1) ν (fun _ m => (m : ℝ)))
          (lossStates A C))
      ∧ HasSum (fun n : lossStates A C => G * lossWeight ν (n : Fin R → ℕ)) 1 := by sorry

end KellyStochasticNetworks
Source
Kelly & Yudovina, Stochastic Networks, CUP 2014, p. 53 (PDF p. 61), equation (3.3): 'If we now truncate n to S(C) = {n : An <= C}, we obtain precisely the original loss network with fixed routing (and finite capacities). Therefore, its equilibrium distribution is pi(n) = G(C) prod_r nu_r^{n_r}/n_r!, n in S(C) = {n : An <= C}, with G(C) = (sum_{n in S(C)} prod_r nu_r^{n_r}/n_r!)^{-1}.' 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).

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