Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 1 (Single-Location Inventory Problem): canonical (s^,S^)(\hat s,\hat S)(s^,S^) optimality, a cost transfer uuu with osc⁡(u)≤K\operatorname{osc}(u)\le Kosc(u)≤K, and the bound TJ−KTJ-KTJ−K

Proved
OWMR.single_location_lemma

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

inventoryinventory-theoryoperations-researchrenewal-theorystochastic-control

Let the one-period demand D≥0D\ge0D≥0 have law ρ\rhoρ on [0,∞)[0,\infty)[0,∞) with 0<ED<∞0<\mathbb E D<\infty0<ED<∞, let f:R→[0,∞)f:\mathbb R\to[0,\infty)f:R→[0,∞) be convex, coercive (f(x)→∞f(x)\to\inftyf(x)→∞ as ∣x∣→∞|x|\to\infty∣x∣→∞) and globally Lipschitz, and fix a setup cost K≥0K\ge0K≥0. In the single-location inventory problem with cost fff, in each period ttt the position may be raised by qt≥0q_t\ge0qt​≥0 at cost K1{qt>0}K\mathbf 1_{\{q_t>0\}}K1{qt​>0}​, the post-action position xtx_txt​ incurs f(xt)f(x_t)f(xt​), and an i.i.d. demand DtD_tDt​ then reduces it. Its value J=JD(f;K)J=J_D(f;K)J=JD​(f;K) (slValue) is the infimum, over all nonanticipative policies that may randomize independently of demand, of the long-run average cost

lim sup⁡T→∞1T E∑t=0T−1[K1{qt>0}+f(xt)]\limsup_{T\to\infty}\frac1T\,\mathbb E\sum_{t=0}^{T-1}\bigl[K\mathbf 1_{\{q_t>0\}}+f(x_t)\bigr]T→∞limsup​T1​Et=0∑T−1​[K1{qt​>0}​+f(xt​)]

from initial position 000. Then JJJ is a finite number with J≥0J\ge0J≥0, and:

  1. Initial position. The same infimum from any initial position y0∈Ry_0\in\mathbb Ry0​∈R equals JJJ.
  2. Optimal (s,S)(s,S)(s,S) rules. If s≤Ss\le Ss≤S and the (s,S)(s,S)(s,S) policy started at SSS has long-run average cost JJJ, then from every initial position the rule that orders up to SSS whenever the pre-action position is at most sss, and does not order otherwise, has long-run average cost JJJ.
  3. Positive setup cost. If K>0K>0K>0, there are s^<S^\hat s<\hat Ss^<S^ such that the (s^,S^)(\hat s,\hat S)(s^,S^) policy started at S^\hat SS^ has long-run average cost JJJ and the canonical conditions hold:
f(s^)=J,f(S^)≤J,f(x)≤J(s^≤x≤S^).f(\hat s)=J,\qquad f(\hat S)\le J,\qquad f(x)\le J\quad(\hat s\le x\le\hat S).f(s^)=J,f(S^)≤J,f(x)≤J(s^≤x≤S^).

Its post-action positions lie in (s^,S^](\hat s,\hat S](s^,S^], and its cycle length τ=min⁡{t≥1: the pre-action position in period t is at most s^}\tau=\min\{t\ge1:\ \text{the pre-action position in period }t\text{ is at most }\hat s\}τ=min{t≥1: the pre-action position in period t is at most s^} has E τm<∞\mathbb E\,\tau^m<\inftyEτm<∞ for every m∈Nm\in\mathbb Nm∈N. 4. Zero setup cost. If K=0K=0K=0, there is a minimizer x∗x^*x∗ of fff with J=f(x∗)=min⁡fJ=f(x^*)=\min fJ=f(x∗)=minf, and base stock at x∗x^*x∗ has long-run average cost JJJ. 5. Cost transfers. There is a measurable function uuu with oscillation sup⁡u−inf⁡u≤K\sup u-\inf u\le Ksupu−infu≤K, and u=0u=0u=0 if K=0K=0K=0, such that

J+(f(x)−J)+≤f(x)+E[u(x−D)−u(x)](x∈R).J+(f(x)-J)^+\le f(x)+\mathbb E\bigl[u(x-D)-u(x)\bigr]\qquad(x\in\mathbb R).J+(f(x)−J)+≤f(x)+E[u(x−D)−u(x)](x∈R).
  1. Finite horizons. Let ξ0,ξ1,…\xi_0,\xi_1,\ldotsξ0​,ξ1​,… be i.i.d. random elements (for example, demand vectors of a network), independent of auxiliary randomness VVV, with Dt=g(ξt)D_t=g(\xi_t)Dt​=g(ξt​) distributed as ρ\rhoρ. For every policy whose order qt≥0q_t\ge0qt​≥0 is a measurable function of VVV and ξ0,…,ξt−1\xi_0,\ldots,\xi_{t-1}ξ0​,…,ξt−1​, and every initial position y0y_0y0​, the post-action positions xt=y0+∑r≤tqr−∑r<tDrx_t=y_0+\sum_{r\le t}q_r-\sum_{r<t}D_rxt​=y0​+∑r≤t​qr​−∑r<t​Dr​ satisfy
TJ−K≤E∑t=0T−1[K1{qt>0}+f(xt)](T≥1).TJ-K\le\mathbb E\sum_{t=0}^{T-1}\bigl[K\mathbf 1_{\{q_t>0\}}+f(x_t)\bigr]\qquad(T\ge1).TJ−K≤Et=0∑T−1​[K1{qt​>0}​+f(xt​)](T≥1).

This is Lemma 1 of the source. It combines the classical average-cost optimality of (s,S)(s,S)(s,S) policies (Iglehart 1963; Zheng 1991; Feinberg and Lewis 2018), here for real positions and every demand law with finite positive mean, with the bounded cost-transfer function and the finite-horizon bound used for the shared-cost lower bound (Proposition 3), where the history may include network demand and auxiliary randomness.

Formalization note. In item 2 the rule is stated recursively on the history: the order in period ttt is S−ytS-y_tS−yt​ if the pre-action position yt=xt−qty_t=x_t-q_tyt​=xt​−qt​ is at most sss, and 000 otherwise; the long-run average cost is that of the deterministic policy. In item 3 the moments are taken under the product law of (D0,D1,…)(D_0,D_1,\ldots)(D0​,D1​,…) (Measure.infinitePi). In item 6 the left side is read as max⁡(TJ−K,0)\max(TJ-K,0)max(TJ−K,0), the right side being nonnegative, and the expectation is over VVV and ξ0,…,ξT−1\xi_0,\ldots,\xi_{T-1}ξ0​,…,ξT−1​.

Preamble
import Definitions.Def_owmr_single_location
import Mathlib.Probability.ProductMeasure

open MeasureTheory
open scoped ENNReal NNReal
Formal statement
namespace OWMR
theorem single_location_lemma (ρ : Measure ℝ≥0) [IsProbabilityMeasure ρ]
    (hρ0 : 0 < ∫⁻ x, (x : ℝ≥0∞) ∂ρ) (hρ1 : ∫⁻ x, (x : ℝ≥0∞) ∂ρ < ∞)
    (f : ℝ → ℝ) (hf0 : ∀ y, 0 ≤ f y) (hfcv : ConvexOn ℝ Set.univ f)
    (hfco : ∀ C : ℝ, ∃ R : ℝ, ∀ y, R ≤ |y| → C ≤ f y) (hfL : ∃ L : ℝ≥0, LipschitzWith L f)
    (K : ℝ) (hK : 0 ≤ K) :
    ∃ J : ℝ, 0 ≤ J ∧ slValue ρ f K = ENNReal.ofReal J ∧
      (∀ y0 : ℝ, (⨅ (U : Type) (_ : MeasurableSpace U) (ν : Measure U)
          (_ : IsProbabilityMeasure ν) (q : SLPolicy U) (_ : SLAdmissible q),
            slAvgCost ρ f K y0 q ν) = slValue ρ f K) ∧
      (∀ s S : ℝ, s ≤ S → ssAvgCost ρ f K s S = slValue ρ f K → ∀ y0 : ℝ,
        ∃ q : SLPolicy Unit, SLAdmissible q ∧
          (∀ (t : ℕ) (d : ℕ → ℝ≥0), q t () d = if slPos y0 q () d t - q t () d ≤ s
            then S - (slPos y0 q () d t - q t () d) else 0) ∧
          slAvgCost ρ f K y0 q (Measure.dirac ()) = slValue ρ f K) ∧
      (0 < K → ∃ s S : ℝ, IsCanonicalOptimum ρ f K s S ∧
        (∀ (d : ℕ → ℝ≥0) (t : ℕ), s < ssPost s S d t ∧ ssPost s S d t ≤ S) ∧
        ∀ m : ℕ, ∫⁻ d, (⨅ (t : ℕ) (_ : 1 ≤ t ∧ ssPre s S d t ≤ s), (t : ℝ≥0∞)) ^ m
          ∂(Measure.infinitePi fun _ : ℕ => ρ) < ∞) ∧
      (K = 0 → ∃ x : ℝ, (∀ y, f x ≤ f y) ∧ J = f x ∧ ssAvgCost ρ f K x x = slValue ρ f K) ∧
      (∃ u : ℝ → ℝ, Measurable u ∧ (∀ y z, u y - u z ≤ K) ∧ (K = 0 → ∀ y, u y = 0) ∧
        ∀ y, J + max (f y - J) 0 ≤ f y + ∫ d, u (y - (d : ℝ)) ∂ρ - u y) ∧
      (∀ {α U : Type} [MeasurableSpace α] [Zero α] [MeasurableSpace U]
        (μ : Measure α) [IsProbabilityMeasure μ] (ν : Measure U) [IsProbabilityMeasure ν]
        (g : α → ℝ≥0), Measurable g → μ.map g = ρ →
        ∀ q : ℕ → U → (ℕ → α) → ℝ,
          (∀ (t : ℕ) (u : U) (d d' : ℕ → α), (∀ r < t, d r = d' r) → q t u d = q t u d') →
          (∀ t, Measurable fun ω : U × (ℕ → α) => q t ω.1 ω.2) → (∀ t u d, 0 ≤ q t u d) →
          ∀ (y0 : ℝ) (T : ℕ), 1 ≤ T →
            ENNReal.ofReal (T * J - K) ≤ ∫⁻ ω, ∑ t ∈ Finset.range T, ENNReal.ofReal
              (K * (if 0 < q t ω.1 (padHist T ω.2) then 1 else 0)
                + f (y0 + ∑ s ∈ Finset.range (t + 1), q s ω.1 (padHist T ω.2)
                  - ∑ s ∈ Finset.range t, (g (padHist T ω.2 s) : ℝ)))
              ∂(ν.prod (Measure.pi fun _ : Fin T => μ))) := 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 4.1 (Single-Location Inventory Problems), Lemma 1 (Single-Location Inventory Problem) with equations (4), (5) and (6), p. 12, and the paragraph before it (value independent of the initial position); proved in Appendix A, pp. E.1-E.3.
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