App. D, p. 35 — stocking J₁ fully gives an adjustable solution of cost 0
ProvedBudgetedAffine.NetworkGap.adjustable_cost_zeroConsider the family of lot-sizing instances (7.2) with nodes, bipartition , , budget uncertainty set , inventory costs on and on , unit capacities, and zero-cost shipping from to . Let the first-stage inventory be
Then there is a second-stage policy such that is feasible for (7.1) and, for every ,
Since every cost is non-negative, this shows : the fully adjustable problem has optimal value zero on the whole family.
Formalization Note The paper constructs (a matching from to ) only for scenarios that are extreme points of . Feasibility requires every , and the statement quantifies over all of ; it remains true, since for fixed the constraints are linear in . Lean indices are 0-based; shipping is restricted to the arcs , the only arcs of finite distance.
import Mathlib import Definitions.Def_BudgetedAffine_NetworkGap_LotSizing import Definitions.Def_BudgetedAffine_NetworkGap_Instance
namespace BudgetedAffine.NetworkGap
/-- App. D, p. 35: with `x = 1` on `J₁` and `x = 0` on `J₂` there is a second-stage policy
`y` such that `(x, y)` is feasible for the family (7.2) and costs `0` in every scenario. -/
theorem adjustable_cost_zero (k : ℕ) :
∃ y : (Node k → ℝ) → Arc k → ℝ,
(lotSizingFamily k).Feasible (Sum.elim (fun _ => 1) (fun _ => 0)) y ∧
∀ h ∈ (lotSizingFamily k).U,
(lotSizingFamily k).cost (Sum.elim (fun _ => 1) (fun _ => 0)) (y h) = 0 := by sorry
end BudgetedAffine.NetworkGap
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.