Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 4 (mirror bounds and candidate cost): B_w≤βw and B_w≤hE[X_w]≤hw/2; a shifted width-w candidate has augmented cost G+(K-∫g(υ+x)U(dx))/M(w)+B_w

Proved
OWMR.mirror_candidate_cost

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

inventoryinventory-theoryoperations-researchrenewal-theorystochastic-control

Fix a retailer iii under the standing assumptions, with F=FiF=F_iF=Fi​, actual setup cost K=KiK=K_iK=Ki​, renewal measure UUU and M(w)=U[0,w)M(w)=U[0,w)M(w)=U[0,w). For w>0w>0w>0 let XwX_wXw​ have the law U(⋅∩[0,w))/M(w)U(\cdot\cap[0,w))/M(w)U(⋅∩[0,w))/M(w) (the long-run depletion of a width-www policy, (10)), and let Bw=min⁡zE[p(z−Xw)++h(Xw−z)+]B_w=\min_z\mathbb E[p(z-X_w)^++h(X_w-z)^+]Bw​=minz​E[p(z−Xw​)++h(Xw​−z)+] be its mirror cost (14), with the minimum written as an infimum. Then:

  1. For every w>0w>0w>0, XwX_wXw​ is a probability law, Bw≤βwB_w\le\beta wBw​≤βw and Bw≤h EXw≤hw/2B_w\le h\,\mathbb EX_w\le hw/2Bw​≤hEXw​≤hw/2, where β=hp/(h+p)\beta=hp/(h+p)β=hp/(h+p).
  2. Let (s^,S^)(\hat s,\hat S)(s^,S^) be a canonical optimum at a planning setup cost k>0k>0k>0, G=J(k)G=J(k)G=J(k) and g(x)=G−F(S^−x)g(x)=G-F(\hat S-x)g(x)=G−F(S^−x). For 0<w≤w^0<w\le\hat w0<w≤w^ and 0≤υ≤w^−w0\le\upsilon\le\hat w-w0≤υ≤w^−w, the width-www policy with target S^−υ\hat S-\upsilonS^−υ (the (s,S)(s,S)(s,S) policy with s=S^−υ−ws=\hat S-\upsilon-ws=S^−υ−w and S=S^−υS=\hat S-\upsilonS=S^−υ), started at its target and charged the actual setup cost KKK, has augmented average cost (average setup, holding and backlog cost plus its mirror cost, augCost) equal to G+(K−∫[0,w)g(υ+x) U(dx))/M(w)+BwG+(K-\int_{[0,w)}g(\upsilon+x)\,U(dx))/M(w)+B_wG+(K−∫[0,w)​g(υ+x)U(dx))/M(w)+Bw​, which is (15).
Preamble
import Definitions.Def_owmr_renewal

open MeasureTheory
open scoped ENNReal NNReal
Formal statement
namespace OWMR
theorem mirror_candidate_cost {n : ℕ} (N : Network n) (hN : N.Standing) (i : Fin n)
    (k sh Sh : ℝ) (hk : 0 < k) (hcan : IsCanonicalOptimum (N.marginal i) (N.F i) k sh Sh)
    (X : ℝ → Measure ℝ) (hX : ∀ w, X w = (renewalMeasure (N.marginal i) (Set.Ico 0 w))⁻¹
      • (renewalMeasure (N.marginal i)).restrict (Set.Ico 0 w))
    (B : ℝ → ℝ)
    (hB : ∀ w, B w = ⨅ z : ℝ, ∫ y, (N.p * max (z - y) 0 + N.h * max (y - z) 0) ∂(X w)) :
    (∀ w, 0 < w → IsProbabilityMeasure (X w) ∧
      B w ≤ N.h * N.p / (N.h + N.p) * w ∧
      B w ≤ N.h * ∫ y, y ∂(X w) ∧ N.h * ∫ y, y ∂(X w) ≤ N.h * w / 2) ∧
    ∃ G : ℝ, 0 < G ∧ N.J i k = ENNReal.ofReal G ∧
      ∀ w υ : ℝ, 0 < w → w ≤ Sh - sh → 0 ≤ υ → υ ≤ Sh - sh - w →
        N.augCost i (Sh - υ - w) (Sh - υ) = ENNReal.ofReal (G
          + (N.K i - ∫ y in Set.Ico 0 w, (G - N.F i (Sh - (υ + y))) ∂(renewalMeasure (N.marginal i)))
            / (renewalMeasure (N.marginal i) (Set.Ico 0 w)).toReal + B w) := by sorry
end OWMR
Source
M. Hu, "A 5/2 Guarantee for Echelon Stock (s,S) Policies in One-Warehouse Multi-Retailer Systems", working paper (manuscript of November 2026), Section 4.3, Lemma 4 (Mirror Bounds and Candidate Cost) with the mirror cost (14), the candidate cost (15), the depletion law of (10) and the augmented cost of Section 3, pp. 13 and 16.
Human review
  • Endorsed by Shuze Chen · Nov 17, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Ming Hu · Nov 17, 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