Theorem 2(b) — the linear-model limit has a smallest minimizer
ProvedSongZipkinFluct.FixedCost.ystar_existsinventorymarkov-decision-processss-policy
Consider the fluctuating-demand inventory model under the standing hypotheses and Assumption 1 (). Let be the limit of the linear order-cost model (fixed cost , terminal cost ), i.e. for the value iteration (10).
Claim (Theorem 2(b)). For every world state , the function on has a smallest global minimizer (it is convex and tends to as ). This is the optimal world-dependent base-stock level of the linear model, and serves as the lower bound in the fixed-cost model.
Preamble
import Mathlib import Definitions.Def_SongZipkinFluct_FixedCost_Bounds
Formal statement
namespace SongZipkinFluct.FixedCost
open Model
/-- Song and Zipkin (1993), Theorem 2(b), p. 357, in the encoding of the fixed-cost model:
under the standing hypotheses and Assumption 1 (`αc̄ < p`), for every world state `i` the limit
`G₀(i, ·) = G_∞(i, ·)` of the linear order-cost model (`K = 0`, `W₀ ≡ 0`) has a smallest global
minimizer `y*(i)`. -/
theorem ystar_exists {I : Type*} [Countable I] [Nonempty I] [DecidableEq I] (M : Model I)
(hM : M.Standing) (hA1 : M.α * M.cbar < M.p) :
∃ ystar, IsYStar M ystar := by
sorry
end SongZipkinFluct.FixedCostSource
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) (y*(i) as used on p. 358)