Equations (2.3) and (2.4) — the partial balance equations
ProvedKellyStochasticNetworks.open_migration_partial_balanceLet be the product-form measure of an open migration process whose solve the traffic equations (2.2). Two families of partial balance equations hold.
- Colony balance, equation (2.3). For every state and every colony with ,
In words: from any state, the rate at which individuals leave colony equals the rate at which they arrive into it.
- Boundary balance, equation (2.4). For every state ,
In words: the rate at which individuals enter the system equals the rate at which they leave it.
Summing the first family over and adding the second gives the equilibrium equations, which is how Theorem 2.8 is proved. Each family is in fact equivalent to the traffic equations: (2.3) reduces to equation (2.2) for the colony , and (2.4) to the sum of (2.2) over all colonies, namely . Partial balance is strictly weaker than detailed balance and strictly stronger than full balance.
Formalization Note The restriction in the first family is the book's implicit requirement that and be states of ; without it, the truncated natural subtraction used for occupancy counts would make the displayed equation assert something the book does not. The term contributes nothing, since .
import Mathlib import Definitions.Def_KellyStochasticNetworks_Balance import Definitions.Def_KellyStochasticNetworks_Migration
namespace KellyStochasticNetworks
theorem open_migration_partial_balance {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)
(htraffic : OpenTraffic lam mu nu α) :
(∀ (n : Fin J → ℕ) (j : Fin J), 1 ≤ n j →
openMigrationPi α g φ n * ((∑ k, lam j k * φ j (n j)) + mu j * φ j (n j))
= (∑ k, openMigrationPi α g φ (Tjk j k n) * (lam k j * φ k (n k + 1)))
+ openMigrationPi α g φ (Tout j n) * nu j)
∧ (∀ n : Fin J → ℕ,
openMigrationPi α g φ n * (∑ k, nu k)
= ∑ k, openMigrationPi α g φ (Tin k n) * (mu k * φ k (n k + 1))) := by sorry
end KellyStochasticNetworksConfirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.
Confirmed by the moderator at approval.