Lemma J.1: service credit (holding and setup cost bound for every policy)
ProvedLostSalesRQ.service_creditapproximation-guaranteeinventorylost-salesoperations-research
Fix , , and a horizon , and write . On every demand environment of rate one, every empty-start admissible policy (lead time ) satisfies
where is on-hand inventory, the number of positive orders and the number of demands served in .
Every policy, however it times and sizes its orders, pays this holding and setup cost per unit served; it is the lower half of the batching tradeoff.
Formalization Note The three expectations are in , so no finiteness hypothesis is needed. "Empty start" means no initial supply.
Preamble
import Definitions.Def_LostSalesRQ_model
Formal statement
namespace LostSalesRQ
theorem service_credit (μ k θ T : ℝ) (hμ : 0 ≤ μ) (hk : 0 ≤ k) (hθ0 : 0 ≤ θ) (hθ1 : θ ≤ 1)
(hT : 0 ≤ T) (S : DemandSpace 1) (π : Policy S) (hπ : π.EmptyStart) :
ENNReal.ofReal (Real.sqrt (2 * k * θ)) * π.expectedServed μ T ≤
ENNReal.ofReal θ * π.expectedHolding μ T + ENNReal.ofReal k * π.expectedOrders T := 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; Lemma J.1, eq. (J.1) (p. A21); main-text statement eq. (25) (p. 22).
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.