Sect. 10.1.1, p. 196: given and , the best feasible is
ProvedInventoryControl.cs_allocation_minclark-scarfinventory-controlmulti-echelonoperations-researchoptimization
Let satisfy the fractile condition (10.8), with , and . In a period in which installation 2 holds the echelon position and the demand over the lead-time turns out to be , installation 1 can realize any position (Eq. 10.1), and among these the cheapest is
If enough stock is available the unconstrained optimum is chosen; otherwise everything available is passed on, because is convex and decreasing to the left of its minimizer. This is the rule an echelon-stock order-up-to- policy implements, and the book stresses that it does not depend on : the optimal policy at installation 1 is determined before anything is known about installation 2.
Preamble
import Definitions.Def_InventoryControl_clarkScarf
Formal statement
namespace InventoryControl
theorem cs_allocation_min (e1 e2 b1 mu sigma : ℝ) (he1 : 0 ≤ e1) (he2 : 0 ≤ e2) (hb : 0 < b1)
(hs : 0 < sigma) (L1 : ℕ) (S1 : ℝ)
(hS1 : ProbabilityTheory.cdf (ProbabilityTheory.gaussianReal 0 1)
((S1 - (L1 + 1) * mu) / (Real.sqrt (L1 + 1) * sigma)) = (e2 + b1) / (e1 + e2 + b1))
(y2 u y1 : ℝ) (hy : y1 ≤ y2 - u) :
csStage1Cost e1 e2 b1 mu sigma L1 (min S1 (y2 - u)) ≤ csStage1Cost e1 e2 b1 mu sigma L1 y1 := by sorry
end InventoryControl
Source
Sven Axsäter, Inventory Control, 3rd ed., Springer 2015, DOI 10.1007/978-3-319-15729-0, p. 196, Sect. 10.1.1: 'If y2 - D(L2) >= y-hat*1 we obtain the optimal solution if we have y1 = y-hat*1. But if y2 - D(L2) < y-hat*1, the best possible value of y1 is y1 = y2 - D(L2) due to the convexity of (10.6)'
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.