§4, (8.0)–(8.T) — a rule in statistical equilibrium yields xᵢⱼ = yᵢq(j | i) satisfying x ≧ 0, (4) and every equation (8.0)–(8.T)
ProvedManneLP.Equilibrium.equilibrium_satisfies_8inventorylinear-programmingmarkov-chainsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Consider Manne's inventory model. Let be a stationary randomized decision rule and a statistical equilibrium of , i.e. a probability vector on with for all . Then the joint probabilities satisfy on the admissible pairs, (4) , and each of the equations
This is the sense in which "equations (8.0)–(8.T) may each be interpreted as a requirement of statistical equilibrium": every equilibrium of every rule gives a feasible point of the linear program.
Formalization Note The equilibrium is defined as a stationary distribution of the chain of , not as the equations (8), so the statement is not a restatement of a definition. No irreducibility is assumed.
Preamble
import Mathlib import Definitions.Def_ManneLP_Equilibrium_Model import Definitions.Def_ManneLP_Equilibrium_Chain import Definitions.Def_ManneLP_Equilibrium_LP
Formal statement
namespace ManneLP.Equilibrium
theorem equilibrium_satisfies_8 (M : Model) (q : ℕ → ℕ → ℝ) (y : ℕ → ℝ)
(hq : IsRule M q) (hy : IsEquilibrium M q y) :
(∀ a ∈ M.A, 0 ≤ jointLaw y q a) ∧ ∑ a ∈ M.A, jointLaw y q a = 1 ∧
Eq80 M (jointLaw y q) ∧ ∀ t : ℕ, 1 ≤ t → t ≤ M.T → Eq8t M (jointLaw y q) t := by sorry
end ManneLP.Equilibrium
Source
Manne, Linear Programming and Sequential Decisions, Management Science 6 (1960), p. 262 (PDF p. 5), §4, (8.0)–(8.T) and the paragraph following them
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.