Erlang loss rate of base stock (M/D/j/j)
ProvedLostSalesRQ.base_stock_served_rateapproximation-guaranteeinventorylost-salesoperations-research
Fix normalized lead-time demand and a demand environment of rate one. For every integer level , every policy following the base-stock rule (, ) serves demand at long-run rate : the expected number of demands served by time , divided by , converges to , where is the Erlang loss fraction.
Under base stock with lead time , units in replenishment form an loss system, and is Erlang's loss probability for it: the statement says that this is the long-run fraction of demand lost.
Formalization Note The statement concerns the expected served count from the stated initialization ( units on hand, empty pipeline), as a limit of time averages.
Preamble
import Definitions.Def_LostSalesRQ_model
Formal statement
namespace LostSalesRQ
theorem base_stock_served_rate (μ : ℝ) (hμ : 0 ≤ μ) (S : DemandSpace 1) (j : ℕ) (hj : 1 ≤ j)
(π : Policy S) (hπ : IsRQPolicy μ (j - 1) 1 π) :
Filter.Tendsto (fun T : ℝ => (π.expectedServed μ T).toReal / T) Filter.atTop
(nhds (1 - lossFraction μ j)) := by sorry
end LostSalesRQSource
M. Hu, "A 1.932 Guarantee for (r,Q) Policies in Continuous-Review Lost-Sales Inventory Systems", arXiv:2610.34576, 2026; Section 3.4 (p. 8) and Appendix A (p. A1), citing Zachary (2007), Theorem 1, p. 241, and Wolff (1982), Theorem 1, p. 225.
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.