Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Corollary 1: an explicit policy selection

Proved
LostSalesRQ.policy_selection

by Ming Hu · Nov 17, 2026 · Mathlib 0df444a (Lean v4.33.1)

approximation-guaranteeinventorylost-salesoperations-research

Suppose h,λ>0h,\lambda>0h,λ>0 and L,p,K≥0L,p,K\ge0L,p,K≥0, and write μ=λL\mu=\lambda Lμ=λL, a=pλ/ha=p\lambda/ha=pλ/h, k=Kλ/hk=K\lambda/hk=Kλ/h. Let S⋆S^\starS⋆ be any cost-minimizing integer base-stock level at zero setup cost, i.e. LS⋆(a)≤Lj(a)L_{S^\star}(a)\le L_j(a)LS⋆​(a)≤Lj​(a) for all j∈Nj\in\mathbb Nj∈N, and set Qk=max⁡{1,⌊2k⌋}Q_k=\max\{1,\lfloor\sqrt{2k}\rfloor\}Qk​=max{1,⌊2k​⌋}, bk=(Qk−1)/2+k/Qkb_k=(Q_k-1)/2+k/Q_kbk​=(Qk​−1)/2+k/Qk​. If S⋆≥1S^\star\ge1S⋆≥1 and Uμ(a)+bk<aU_\mu(a)+b_k<aUμ​(a)+bk​<a, choose the (r,Q)=(S⋆−1,Qk)(r,Q)=(S^\star-1,Q_k)(r,Q)=(S⋆−1,Qk​) policy; otherwise choose no replenishment. Then:

  1. on every demand environment of rate λ\lambdaλ, such a policy exists;
  2. every selected policy π\piπ satisfies
C(π)≤R OPT,R={2,K=0,(2+6)/2,in general;C(\pi)\le R\,\mathrm{OPT},\qquad R=\begin{cases}\sqrt2,&K=0,\\ (\sqrt2+\sqrt6)/2,&\text{in general};\end{cases}C(π)≤ROPT,R={2​,(2​+6​)/2,​K=0,in general;​
  1. when K=0K=0K=0, π\piπ is an integer base-stock policy (including level zero) and is optimal among integer base-stock policies: C(π)≤C(π′)C(\pi)\le C(\pi')C(π)≤C(π′) for every π′∈ΠS\pi'\in\Pi_{\rm S}π′∈ΠS​.

The corollary turns the guarantee into a finite procedure that needs neither OPT\mathrm{OPT}OPT nor any optimization over QQQ.

Formalization Note The selected policy is described by the predicate IsSelectedPolicy; its existence is part of the statement, so the conclusions are not vacuous. For K=0K=0K=0 the stated factor is 2\sqrt22​; for K>0K>0K>0 it is (2+6)/2(\sqrt2+\sqrt6)/2(2​+6​)/2.

Preamble
import Definitions.Def_LostSalesRQ_model
Formal statement
namespace LostSalesRQ

theorem policy_selection (lam L h p K : ℝ) (hlam : 0 < lam) (hL : 0 ≤ L) (hh : 0 < h)
    (hp : 0 ≤ p) (hK : 0 ≤ K) (S : DemandSpace lam) (Sstar : ℕ)
    (hS : ∀ j : ℕ, baseStockCost (lam * L) (p * lam / h) Sstar ≤
      baseStockCost (lam * L) (p * lam / h) j) :
    (∃ π : Policy S, IsSelectedPolicy L (lam * L) (p * lam / h) (K * lam / h) Sstar π) ∧
    ∀ π : Policy S, IsSelectedPolicy L (lam * L) (p * lam / h) (K * lam / h) Sstar π →
      π.avgCost L h p K ≤
          ENNReal.ofReal (if K = 0 then Real.sqrt 2 else (Real.sqrt 2 + Real.sqrt 6) / 2) *
            OPT lam L h p K ∧
        (K = 0 → π ∈ baseStockFamily S L ∧
          ∀ π' ∈ baseStockFamily S L, π.avgCost L h p 0 ≤ π'.avgCost L h p 0) := by sorry

end LostSalesRQ
Source
M. Hu, "A 1.932 Guarantee for (r,Q) Policies in Continuous-Review Lost-Sales Inventory Systems", arXiv:2610.34576, 2026; Corollary 1 (p. 9).
Human review
  • Endorsed by Shuze Chen · Nov 18, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Ming Hu · Nov 18, 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