Section 3.3, p. 55 — ΔB(s) = −(1 − Σ_{x≤s} p(x)), Δ²B(s) = p(s+1): expected backorders are decreasing and discretely convex
ProvedServiceParts.StockLevels.backorder_differencesLet an item have compound Poisson demand with steady-state resupply probabilities , ready rate and expected backorders . Then for every stock level
so is discretely convex. If demand is simple Poisson (), then moreover
that is, is strictly decreasing and strictly discretely convex.
Convexity of is the property that makes backorder-based stocking problems tractable, in contrast with the fill rate and the ready rate.
Formalization Note The book derives these identities in Section 3.3 under simple Poisson demand, and uses the difference identity for compound Poisson demand on p. 61. The two identities and weak convexity are stated for compound Poisson demand (a labelled generalization); the strict statements are stated for simple Poisson demand, as in the book (for compound demand can vanish).
import Mathlib import Definitions.Def_ServiceParts_StockLevels_Basic import Definitions.Def_ServiceParts_StockLevels_CompoundPoissonDemand
namespace ServiceParts.StockLevels
theorem backorder_differences (d : CompoundPoissonDemand) (s : ℕ) :
fdiff d.backorders s = -(1 - d.readyRate s) ∧
fdiff2 d.backorders s = d.pmf (s + 1) ∧
0 ≤ fdiff2 d.backorders s ∧
(d.u 1 = 1 → fdiff d.backorders s < 0 ∧ 0 < fdiff2 d.backorders s) := by sorry
end ServiceParts.StockLevels
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.