Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2 — infinite-horizon world-dependent basestock optimality

Proved
SongZipkinFluct.Linear.theorem2

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

dynamic-programminginventorymarkov-modulated-demandp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

In the linear-order-cost model, under αcˉ<p\alpha\bar c<pαcˉ<p, every world state iii has a finite smallest global minimizer y∗(i)y^*(i)y∗(i) of the limiting cost G∞(i,⋅)G_\infty(i,\cdot)G∞​(i,⋅). The finite-stage minimizers converge to a level y∞∗(i)y_\infty^*(i)y∞∗​(i), and the myopic, limiting and optimal levels satisfy

0≤ymin⁡+≤y∗(i)≤y∞∗(i)≤y+(i).0\le y^+_{\min}\le y^*(i)\le y_\infty^*(i)\le y^+(i).0≤ymin+​≤y∗(i)≤y∞∗​(i)≤y+(i).

The level y∞∗(i)y_\infty^*(i)y∞∗​(i) also minimizes G∞(i,⋅)G_\infty(i,\cdot)G∞​(i,⋅). The stationary world-dependent basestock policy that orders to max⁡{x,y∗(i)}\max\{x,y^*(i)\}max{x,y∗(i)} in state (i,x)(i,x)(i,x) minimizes infinite-horizon expected discounted cost from every state among all feasible history-dependent policies.

Formalization Note The policy-cost recursion uses the uniformized version of equation (1), with nonnegative extended-real costs. Its policy comparison includes every feasible deterministic history-dependent policy.

The cost of this basestock policy is finite from every initial state; this prevents an all-infinite-cost reading of optimality.

Preamble
import Mathlib
import Definitions.Def_SongZipkinFluct_Linear_Dynamics
Formal statement
namespace SongZipkinFluct.Linear

/-- Song and Zipkin (1993), Theorem 2(b), (c), (e), pp. 357--358.
The existence of the finite smallest minimizers and their limiting
integer levels is asserted, not assumed. The basestock policy is
optimal from every initial state among all feasible deterministic
history-dependent policies of the uniformized infinite-horizon MDP
of equation (1). Randomization is omitted because it cannot improve
nonnegative expected costs in this countable-state model. -/
theorem theorem2 {I : Type*} [Countable I] [Nonempty I] [DecidableEq I]
    (M : Model I) (hA : M.Assumption1) :
    ∃ (ys : ℕ → I → ℤ) (yplus ystar yinf : I → ℤ) (ymin : ℤ),
      (∀ i : I, SmallestMinimizer (M.Gplus i) (yplus i)) ∧
      IsLeast (Set.range yplus) ymin ∧
      (∀ n : ℕ, 1 ≤ n → ∀ i : I,
        SmallestMinimizer (M.Glin n i) (ys n i)) ∧
      (∀ i : I, Filter.Tendsto (fun n : ℕ => ys (n + 1) i)
        Filter.atTop (nhds (yinf i))) ∧
      (∀ i : I, SmallestMinimizer (M.Ginf i) (ystar i)) ∧
      0 ≤ ymin ∧
      (∀ i : I,
        ymin ≤ ystar i ∧ ystar i ≤ yinf i ∧ yinf i ≤ yplus i ∧
        ∀ y : ℤ, M.Ginf i (yinf i) ≤ M.Ginf i y) ∧
      M.IsOptimal 0 (baseStock ystar) ∧
      (∀ s : State I, M.policyCost 0 (baseStock ystar) [] s < ⊤) := by sorry

end SongZipkinFluct.Linear
Source
Song and Zipkin, Inventory Control in a Fluctuating Demand Environment, Oper. Res. 41(2):351–370 (1993), DOI 10.1287/opre.41.2.351, pp. 357–358, Theorem 2(b), (c), (e)
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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