§2, p. 1257 — the one-period costs satisfy q^l(x^l) = γE[q^{l−1}(x_+^{l−1})]
ProvedLostSalesOldNew.LeadTime.q_recursionConsider the lost-sales system of Zipkin (2008), §2, with cost data , discount factor , transformed holding cost , and i.i.d. demand with law on with finite mean. Recall and , where is the stock on hand periods after the state .
For every and every ,
where and the expectation is over the current period's demand .
The recursion computes the transformed one-period cost one period at a time, and it is the identity that makes the terms cancel in the paper's derivation of the lead-time bound (6).
Formalization Note. The paper states the recursion for ; it is stated here for every (in Lean, l + 1 for l : ℕ), which contains the paper's range. The state is Fin (l + 2) → ℝ and need not be nonnegative. Finite mean demand (Integrable id D) is assumed so that the Bochner integrals defining and are the paper's expectations.
import Mathlib import Definitions.Def_LostSalesOldNew_LeadTime_Model
namespace LostSalesOldNew.LeadTime
open MeasureTheory
theorem q_recursion (K : Costs) (D : Measure ℝ) [IsProbabilityMeasure D]
(hD : LostSalesOldNew.StateReduction.IsDemand D) (hmean : Integrable (fun d : ℝ => d) D)
(l : ℕ) (x : Fin (l + 2) → ℝ) :
qL K D (l + 1) x = K.γ * ∫ d, qL K D l (LostSalesOldNew.StateReduction.shift x d) ∂D := by sorry
end LostSalesOldNew.LeadTime
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.