Section 3.4: exact zero-setup costs of the base-stock policies
ProvedLostSalesRQ.exact_base_stock_costFix normalized lead-time demand , penalty (unit demand rate, unit holding cost, no setup cost) and any demand environment of rate one. The empty no-order policy has average cost , and for every integer level every policy following the base-stock rule (, ) has average cost
Pending replacements form an loss system, whose occupied-position count is Poisson conditioned to be at most ; is the long-run fraction of demand lost and the mean on-hand inventory. These closed forms are what the selection algorithm minimizes.
Formalization Note The average cost is the upper limit of expected average costs from the stated initialization ( units on hand, empty pipeline), not a stationary expectation; the statement asserts that it equals the Erlang formula.
import Definitions.Def_LostSalesRQ_model
namespace LostSalesRQ
theorem exact_base_stock_cost (μ a : ℝ) (hμ : 0 ≤ μ) (ha : 0 ≤ a) (S : DemandSpace 1) :
(noOrderPolicy S).avgCost μ 1 a 0 = ENNReal.ofReal a ∧
∀ j : ℕ, 1 ≤ j → ∀ π : Policy S, IsRQPolicy μ (j - 1) 1 π →
π.avgCost μ 1 a 0 = ENNReal.ofReal (baseStockCost μ a j) := by sorry
end LostSalesRQConfirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.