Theorem 2(b): pointwise and has a smallest minimizer
OpenSongZipkinFluct.Monotone.theorem2_bbasestock-policyinventorymarkov-modulated-demand
Assume the standing hypotheses of the model and Assumption 1 (), in the linear-cost model (, terminal cost ) with the recursion (13):
Let . Then
- for every and , as (so the supremum is a genuine limit), and
- for every , has a smallest minimizer over .
This is the part of Theorem 2(b) of the paper (p. 357) stating that converges to , which is convex and coercive in , so that its smallest minimizer exists.
Preamble
import Mathlib import Definitions.Def_SongZipkinFluct_Monotone_Recursion open Filter Topology
Formal statement
namespace SongZipkinFluct.Monotone
theorem theorem2_b {I : Type} [Countable I] [Nonempty I] [DecidableEq I]
(M : Model I) (hM : M.Standing) (hA : M.Assumption1) :
(∀ (i : I) (y : ℤ),
Tendsto (fun n : ℕ => M.Gn 0 0 (n + 1) i y) atTop (𝓝 (M.Ginf 0 0 i y))) ∧
∃ ystar : I → ℤ, ∀ i, IsSmallestMinimizer (M.Ginf 0 0 i) (ystar i) := by sorry
end SongZipkinFluct.Monotone
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, p. 357, Theorem 2(b) (with Theorem 1 (printed "Theorem 4") (d)–(e) giving monotone convergence)