The equilibrium distribution of the Erlang loss link
ProvedKellyStochasticNetworks.erlang_link_equilibriumConsider the Erlang loss link with parallel circuits, calls arriving at rate and holding times exponential of parameter , whose transition rates on the state space are
If a collection of numbers satisfies the detailed balance equations for these rates, then for every state
This is the solution of the detailed balance recursion on p. 18 of the book. It determines the equilibrium distribution up to the single unknown , which the normalization then fixes. No normalization is assumed here, so the statement applies to any solution of the detailed balance equations, normalized or not.
Formalization Note The state space is Fin (C + 1) and the exponent is the underlying
natural number of that index, so the factorial is the ordinary factorial of a natural number.
import Mathlib import Definitions.Def_KellyStochasticNetworks_Balance import Definitions.Def_KellyStochasticNetworks_Erlang
namespace KellyStochasticNetworks
theorem erlang_link_equilibrium (lam mu : ℝ) (C : ℕ) (hlam : 0 < lam) (hmu : 0 < mu)
(π : Fin (C + 1) → ℝ) (h : DetailedBalance π (erlangRates lam mu C)) (j : Fin (C + 1)) :
π j = (lam / mu) ^ (j : ℕ) / (Nat.factorial (j : ℕ) : ℝ) * π 0 := by sorry
end KellyStochasticNetworksConfirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.