Lemma 6.1: is convex in and minimized at the multiple of lying in
ProvedInventoryControl.lemma_6_1discrete-optimizationinventory-controloperations-researchreorder-point
Let be the holding-plus-shortage cost rate of a discrete lead-time demand with finite mean, with costs and , let , let , and let be an integer minimizing . Then for every integer :
- is convex on , in the sense of nondecreasing increments;
- there is a unique integer with , and the element
of the band is
reduceToBand R Q z; - for every integer .
The proof rests on the telescoping identity (6.21), , which is nonpositive while and nonnegative once , because minimizes the convex function . The lemma is the pointwise comparison behind Proposition 6.1: whatever inventory position a policy holds, the position in the band congruent to it modulo costs no more.
Formalization Note Part 2 is stated as three facts about reduceToBand R Q z: it lies in
the band and differs from by a multiple of ; uniqueness is immediate from the band's
length.
Preamble
import Definitions.Def_InventoryControl_rqPolicy
Formal statement
namespace InventoryControl
theorem lemma_6_1 (D : DiscreteDemand) (h b1 : ℝ) (hh : 0 < h) (hb : 0 < b1) (Q : ℕ) (hQ : 0 < Q)
(R : ℤ) (hR : ∀ y : ℤ, windowCost D h b1 Q R ≤ windowCost D h b1 Q y) (z : ℤ) :
(∀ x : ℤ, sPolicyCost D h b1 (z + (x + 1) * Q) - sPolicyCost D h b1 (z + x * Q)
≤ sPolicyCost D h b1 (z + (x + 2) * Q) - sPolicyCost D h b1 (z + (x + 1) * Q))
∧ (R + 1 ≤ reduceToBand R Q z ∧ reduceToBand R Q z ≤ R + Q
∧ ∃ x : ℤ, reduceToBand R Q z = z + x * Q)
∧ ∀ x : ℤ, sPolicyCost D h b1 (reduceToBand R Q z) ≤ sPolicyCost D h b1 (z + x * Q) := by sorry
end InventoryControl
Source
Sven Axsäter, Inventory Control, 3rd ed., Springer 2015, DOI 10.1007/978-3-319-15729-0, p. 114, Sect. 6.2.1, Lemma 6.1 and its proof, Eq. (6.21)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.