Proof of Theorem 4, p. 939 — if f̄_{t+1} is L♮-convex, so is ḡ_t(v, ζ) of (5)
ProvedZipkinLostSales.LNatural.gbar_lnatconvexConsider the lost-sales model with lead time , nonnegative independent demands with common law of finite mean, nonnegative unit costs and discount factor . Let and be the optimal cost functions with periods to go in the transformed state, so that
If (the paper's ) is L♮-convex on , then (the paper's ) is L♮-convex on .
In the paper this is the step of (5), with L♮-convexity preserved by expectation; together with Lemma 2 it closes the induction of Theorem 4.
Formalization Note. is defined by the recursion (1) in the transformed state, not by (5); the two agree by the reformulation milestone (optimal ). Time is the number of periods to go. The cost hypotheses read "unit cost" and "discount rate"; finite mean demand is added so that is finite. Expectations are Bochner integrals against .
import Mathlib import Definitions.Def_ZipkinLostSales_LNatural_Model open MeasureTheory
namespace ZipkinLostSales.LNatural
/-- Proof of Theorem 4 (Zipkin 2008, p. 939): if `f̄_{t+1}` (here `fbar … k`) is L♮-convex on `V`,
then `ḡ_t(v, ζ) = −γ^L c ζ + E_d[κ̄_t(v, ζ | d)]` (here `gbar … k`) is L♮-convex on `V × ℜ⁻`. -/
theorem gbar_lnatconvex (L : ℕ) (hL : 0 < L) (μ : Measure ℝ) [IsProbabilityMeasure μ]
(hμ : μ (Set.Iio 0) = 0) (hμint : Integrable id μ)
(c hh p γ : ℝ) (hc : 0 ≤ c) (hhh : 0 ≤ hh) (hp : 0 ≤ p) (hγ0 : 0 < γ) (hγ1 : γ ≤ 1)
(k : ℕ) (hk : LNatConvexOn (V L) (fbar c hh p γ μ k)) :
LNatConvexOn (VxNeg L) (liftG (gbar c hh p γ μ k)) := by sorry
end ZipkinLostSales.LNatural
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.