(12)–(13) — the SD problem decomposes into N single-product newsvendor problems
ProvedMixFlex.RiskNeutral.sd_optimal_investmentflexibilitynewsvendorp2o-batch-p200ap2o-gran-per-chapterp2o-plan-paperp2o-v1unreliable-supply
In the SD network of the mission's model, suppose and let satisfy, for every product ,
Then maximizes over all , and the optimal SD value is
These are equations (12)–(13): the SD investment problem splits into independent single-product problems, each solved by the same critical fractile.
Formalization Note As in (10)–(11), is replaced by the hypothesis fractile, and by . is needed for the fractile to be defined.
Preamble
import Mathlib import Definitions.Def_MixFlex_RiskNeutral_Model
Formal statement
namespace MixFlex.RiskNeutral
open MeasureTheory ProbabilityTheory
/-- (12)–(13), p. 42: the SD problem decomposes into `N` single-product problems; any `K* ≥ 0`
with `F_{X_n}(K*_n) = 1 − (λ + (1 − λ)θ) c / (θ p)` for every `n` is an optimal SD investment,
and the optimal SD value is `θ p Σ_n E[X_n · 1{X_n ≤ K*_n}]`. -/
theorem sd_optimal_investment (P : Params) (hP : P.Standing) {N : ℕ} {Ω : Type*}
[MeasurableSpace Ω] (μ : Measure Ω) (X : Ω → Fin N → ℝ) (Y : Ω → Fin (N + 1) → ℝ)
(hS : Setting P μ X Y) (htheta : 0 < P.theta) (Kstar : Fin N → ℝ) (hK : ∀ n, 0 ≤ Kstar n)
(hq : ∀ n, μ.real {ω | X ω n ≤ Kstar n}
= 1 - (P.lam + (1 - P.lam) * P.theta) * P.c / (P.theta * P.p)) :
(∀ K : Fin N → ℝ, (∀ n, 0 ≤ K n) → vSD P μ X Y K ≤ vSD P μ X Y Kstar) ∧
vSD P μ X Y Kstar
= P.theta * P.p * ∑ n, ∫ ω, X ω n * (if X ω n ≤ Kstar n then 1 else 0) ∂μ := by sorry
end MixFlex.RiskNeutral
Source
Tomlin and Wang, On the value of mix flexibility and dual sourcing in unreliable newsvendor networks, Manufacturing Service Oper. Management 7(1), 2005, p. 42, §3.1, (12)–(13)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.