Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Section 3.3, p. 53 — ΔF(s) = p(s|λτ̄), Δ²F(s) = p(s|λτ̄)(λτ̄/(s+1) − 1): the Poisson fill rate is concave exactly from ⌊λτ̄⌋ on

Proved
ServiceParts.StockLevels.fill_rate_differences

by mikedeng1 · Oct 2, 2026 · Mathlib 0df444a (Lean v4.33.1)

discrete-convexityfill-rateinventoryp2o-batch-b23ap2o-gran-per-chapterp2o-plan-bookp2o-v1poisson-distribution

Let demand be a simple Poisson process and let a=λτˉ>0a = \lambda\bar\tau > 0a=λτˉ>0 be the mean number of units in resupply, so that p(x∣a)=e−aax/x!p(x \mid a) = e^{-a}a^x/x!p(x∣a)=e−aax/x!. The fill rate at stock level sss is F(s)=∑x<sp(x∣a)F(s) = \sum_{x < s} p(x \mid a)F(s)=∑x<s​p(x∣a). Then for every s≥0s \ge 0s≥0:

  1. ΔF(s)=F(s+1)−F(s)=p(s∣a)=e−a as/s!\Delta F(s) = F(s+1) - F(s) = p(s \mid a) = e^{-a}\,a^s/s!ΔF(s)=F(s+1)−F(s)=p(s∣a)=e−aas/s!;
Δ2F(s)=e−aas+1(s+1)!−e−aass!=e−aass!(as+1−1);\Delta^2 F(s) = e^{-a}\frac{a^{s+1}}{(s+1)!} - e^{-a}\frac{a^{s}}{s!} = e^{-a}\frac{a^s}{s!}\Big(\frac{a}{s+1} - 1\Big);Δ2F(s)=e−a(s+1)!as+1​−e−as!as​=e−as!as​(s+1a​−1);
  1. if a>s+1a > s + 1a>s+1 then Δ2F(s)>0\Delta^2 F(s) > 0Δ2F(s)>0 (the fill rate is not concave there);
  2. if aaa is not an integer, Δ2F(s)≤0\Delta^2 F(s) \le 0Δ2F(s)≤0 if and only if s≥⌊a⌋s \ge \lfloor a \rfloors≥⌊a⌋;
  3. if aaa is an integer, Δ2F(s)≤0\Delta^2 F(s) \le 0Δ2F(s)≤0 if and only if s≥a−1s \ge a - 1s≥a−1.

This is why stock levels in fill-rate optimization problems are restricted to s≥⌊λτˉ⌋s \ge \lfloor\lambda\bar\tau\rfloors≥⌊λτˉ⌋: only there is the fill rate discretely concave.

Formalization Note λ\lambdaλ and τˉ\bar\tauτˉ are positive reals and a=λτˉa = \lambda\bar\taua=λτˉ. "Integer" means a=na = na=n for a natural number nnn, which is the only possibility since a>0a > 0a>0.

Preamble
import Mathlib
import Definitions.Def_ServiceParts_StockLevels_Basic
Formal statement
namespace ServiceParts.StockLevels

theorem fill_rate_differences (lam tbar : ℝ) (hlam : 0 < lam) (htbar : 0 < tbar) (s : ℕ) :
    fdiff (poissonFillRate (lam * tbar)) s = poissonPmf (lam * tbar) s ∧
    fdiff2 (poissonFillRate (lam * tbar)) s
      = poissonPmf (lam * tbar) (s + 1) - poissonPmf (lam * tbar) s ∧
    fdiff2 (poissonFillRate (lam * tbar)) s
      = poissonPmf (lam * tbar) s * (lam * tbar / ((s : ℝ) + 1) - 1) ∧
    ((s : ℝ) + 1 < lam * tbar → 0 < fdiff2 (poissonFillRate (lam * tbar)) s) ∧
    ((¬ ∃ n : ℕ, lam * tbar = n) →
      (fdiff2 (poissonFillRate (lam * tbar)) s ≤ 0 ↔ ⌊lam * tbar⌋₊ ≤ s)) ∧
    ((∃ n : ℕ, lam * tbar = n) →
      (fdiff2 (poissonFillRate (lam * tbar)) s ≤ 0 ↔ lam * tbar - 1 ≤ (s : ℝ))) := by sorry

end ServiceParts.StockLevels
Source
Muckstadt, Analysis and Algorithms for Service Parts Supply Chains, Springer 2005, DOI 10.1007/b138879, pp. 52-53, Section 3.3 (F(s), ΔF(s), Δ²F(s) and the concavity region)
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me