Lemma 3, p. 164 — moving ε of demand from class i to a less important class j does not raise f_t or g_t
ProvedTopkisRation.Levels.lemma_3dynamic-programminginventoryp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1rationing
In the model of Topkis (1968, §1) under Assumptions (A)–(C), let be two demand classes, the corresponding unit vectors, and . Then:
- for , and with ,
- for , and with ,
Demand of a more important class is never cheaper to face than the same amount of a less important class. The lemma relies on the backlog relation , and it is used to show that the priority vector (5) is optimal (Lemma 4).
Formalization Note. The page states the single hypothesis for both displays; each display carries here the part that concerns its own vector, which is the same claim.
Preamble
import Mathlib import Definitions.Def_TopkisRation_Levels_Model
Formal statement
namespace TopkisRation.Levels
theorem lemma_3 {n : ℕ} (M : Model n) (hM : M.Standing) (i j : Fin n) (hij : j < i)
(ε : ℝ) (hε : 0 < ε) :
(∀ t ∈ Finset.Icc 1 M.k, ∀ z, 0 ≤ z → ∀ B : Fin n → ℝ, 0 ≤ B → ε ≤ B i →
M.f t z (B - ε • (Pi.single i 1 - Pi.single j 1)) ≤ M.f t z B) ∧
(∀ t ≤ M.k, ∀ z, 0 ≤ z → ∀ b : Fin n → ℝ, 0 ≤ b → ε ≤ b i →
M.g t z (b - ε • (Pi.single i 1 - Pi.single j 1)) ≤ M.g t z b) := by sorry
end TopkisRation.Levels
Source
Topkis, Optimal ordering and rationing policies in a nonstationary dynamic inventory model with n demand classes, Management Science 15 (1968), p. 164, Lemma 3, (2) and (3)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.