Theorem 6.1, (6.8), p. 937 — 0 = v_{n,0} ≤ v_{n,1} ≤ ⋯ ≤ v_{n,k} ≤ w_n
ProvedSethiChengSS.Infinite.theorem_6_1_eq_6_8dynamic-programminginfinite-horizoninventoryp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1value-iteration
Under the standing assumptions (2.1)–(2.2) and , the values of the -period truncations (6.4) increase in and are bounded by the order-nothing cost of (6.6):
Monotonicity and the bound make the limit exist, which is the first step of the successive approximation of the infinite-horizon problem.
Formalization Note. Values are in . The paper allows ; is assumed throughout the mission (the Appendix divides by ).
Preamble
import Mathlib import Definitions.Def_BertsekasKConvex import Definitions.Def_SethiChengSS_Infinite_Model
Formal statement
namespace SethiChengSS.Infinite
theorem theorem_6_1_eq_6_8 {L : ℕ} (D : Model L) (α : ℝ) (hD : Standing D) (hα0 : 0 < α)
(hα1 : α < 1) :
(∀ n i x, vTrunc D α n 0 i x = 0) ∧
(∀ n k i x, vTrunc D α n k i x ≤ vTrunc D α n (k + 1) i x) ∧
(∀ n k i x, vTrunc D α n k i x ≤ w D α n i x) := by sorry
end SethiChengSS.Infinite
Source
Sethi and Cheng, Optimality of (s, S) policies in inventory models with Markovian demand, Oper. Res. 45(6) (1997), DOI 10.1287/opre.45.6.931, p. 937, Theorem 6.1, (6.8)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.