Theorem 6.3 (Clark-Scarf): the sequentially optimized echelon base-stock vector is optimal, with cost
DisprovedSupplyChainTheory.clark_scarf_sequentialTheorem 6.3. Consider an -stage serial system with echelon holding costs , stockout cost at stage 1 and lead-time demands of finite mean. Let and, for ,
Then is the optimal echelon base-stock vector and is the corresponding optimal cost: for every echelon base-stock vector , whose expected cost is computed by the same recursion (6.21)-(6.23) with in place of ,
Rather than optimizing all base-stock levels simultaneously, one optimizes them one stage at a time from the customer upward, each step a single-variable convex minimization. Zipkin calls (6.24)-(6.27) the fundamental equations of supply chain theory; the result is Clark and Scarf's (1960) for finite horizons and Federgruen and Zipkin's and Chen and Zheng's for the infinite-horizon form used here.
Formalization Note The theorem is about the recursion: the identification of
with the steady-state expected cost of the physical system is the book's
derivation (6.12)-(6.20), which is not restated. enters as any vector for which each
minimizes ; such a vector exists under ,
(cs_sequential_exists).
Retired. Retired because N = 0 was admissible: CSSequential 0 … is vacuous and the conclusion compares csG … 0 at the unrelated points S*₀ and S₀ (accepted disproof); replaced by SupplyChainTheory.clark_scarf_sequential_v2 with 1 ≤ N.
import Definitions.Def_SupplyChainTheory_multiechelon
namespace SupplyChainTheory
theorem clark_scarf_sequential (N : ℕ) (h : ℕ → ℝ) (p : ℝ) (D : ℕ → MeasureTheory.Measure ℝ)
[∀ j, MeasureTheory.IsProbabilityMeasure (D j)]
(hD : ∀ j, MeasureTheory.Integrable (fun x => x) (D j))
(hh : ∀ j, 0 ≤ h j) (hp : 0 ≤ p) (Sstar : ℕ → ℝ) (hS : CSSequential N h p D Sstar) :
∀ S : ℕ → ℝ, csG N h p D Sstar N (Sstar N) ≤ csG N h p D S N (S N) := by sorry
end SupplyChainTheory
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.