Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Corollary 2 (p. 201) — (t+1)−1P[δ0t]⋯P[δtt](Hιv1+HιG)→0(t+1)^{-1}P[\delta_{0t}]\cdots P[\delta_{tt}](H_\iota v1 + H_\iota G) \to 0(t+1)−1P[δ0t​]⋯P[δtt​](Hι​v1+Hι​G)→0

Proved
FedergruenZipkin.AvgCost.corollary2_cesaro_vanish

by mikedeng1 · Oct 4, 2026 · Mathlib 0df444a (Lean v4.33.1)

inventorymarkov-chainp2o-batch-pfp1bp2o-gran-per-chapterp2o-plan-paperp2o-v1

In the capacitated inventory model with storage capacity U≥yˉ∞U \ge \bar y^\inftyU≥yˉ​∞, where yˉ∞\bar y^\inftyyˉ​∞ is the smallest global minimizer of GGG, set L=yˉ∞−bL = \bar y^\infty - bL=yˉ​∞−b. For a policy δ\deltaδ let P[δ]P[\delta]P[δ] be the transition operator P[δ]w(x)=E w(δ(x)−D)P[\delta]w(x) = E\,w(\delta(x) - D)P[δ]w(x)=Ew(δ(x)−D). Let ι=[l,u]\iota = [l, u]ι=[l,u] satisfy the assumptions of Lemma 3 (0≤u≤U0 \le u \le U0≤u≤U, l≤u−bl \le u - bl≤u−b), and let {(δ0t,δ1t,…,δtt)}t≥0\{(\delta_{0t}, \delta_{1t}, \dots, \delta_{tt})\}_{t \ge 0}{(δ0t​,δ1t​,…,δtt​)}t≥0​ be any sequence of (t+1)(t+1)(t+1)-period policies with every δst∈ΔL\delta_{st} \in \Delta_Lδst​∈ΔL​ (feasible and ordering to capacity at or below LLL). Then for all x≤Ux \le Ux≤U

lim⁡t→∞(t+1)−1 P[δ0t]P[δ1t]⋯P[δtt](Hιv1+HιG)(x)=0.\lim_{t\to\infty} (t+1)^{-1}\, P[\delta_{0t}]P[\delta_{1t}]\cdots P[\delta_{tt}]\bigl(H_\iota v1 + H_\iota G\bigr)(x) = 0.t→∞lim​(t+1)−1P[δ0t​]P[δ1t​]⋯P[δtt​](Hι​v1+Hι​G)(x)=0.

This is the condition that lets the paper pass from the optimality equation to strong optimality in Theorem 1(b): the expected relative value at time ttt grows sublinearly under any sequence of policies in ΔL\Delta_LΔL​.

Formalization Note The corollary leaves ι\iotaι implicit; it is stated here for every ι\iotaι satisfying the assumptions of Lemma 3, as in Corollary 1 (the proof of Theorem 1(b) uses ι=[L,U]\iota = [L, U]ι=[L,U]). The sequence of policies is a function σ(t,s)=δst\sigma(t, s) = \delta_{st}σ(t,s)=δst​, constrained only for s≤ts \le ts≤t. Values are in [0,∞][0, \infty][0,∞].

Preamble
import Mathlib
import Definitions.Def_FedergruenZipkin_AvgCost_Model
Formal statement
namespace FedergruenZipkin.AvgCost
open scoped ENNReal
open Filter Topology
theorem corollary2_cesaro_vanish (M : Model) (U yInf : ℤ)
    (hyInf : IsLeast {y : ℤ | ∀ z, M.G y ≤ M.G z} yInf) (hU : yInf ≤ U)
    (l u : ℤ) (hu0 : 0 ≤ u) (huU : u ≤ U) (hlu : l ≤ u - M.b)
    (σ : ℕ → ℕ → ℤ → ℤ) (hσ : ∀ t s, s ≤ t → FeasibleL M U (yInf - M.b) (σ t s)) :
    ∀ x ≤ U, Tendsto (fun t : ℕ =>
        Pcomp M (σ t) t (H M U l u (fun _ => (1 : ℝ≥0∞)) + H M U l u (fun y => ENNReal.ofReal (M.G y))) x /
          ((t : ℝ≥0∞) + 1)) atTop (𝓝 0) := by sorry
end FedergruenZipkin.AvgCost
Source
Federgruen and Zipkin, An Inventory Model with Limited Production Capacity and Uncertain Demands I, Math. Oper. Res. 11(2), 1986, pp. 200-201, Corollary 2 (with the definition of P[δ] and L preceding it)
Human review
  • Endorsed by Shuze Chen · Oct 4, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 4, 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