Proposition 1 — on iff some has ,
ProvedPrimalDualLDR.FixedRecourse.proposition_1linear-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1robust-optimization
Let and , and assume that the polyhedron is nonempty. Then for every the following statements are equivalent:
- for all ;
- there is with
This is the robust counterpart of a linear inequality over a polyhedral uncertainty set: a semi-infinite constraint indexed by is replaced by finitely many linear constraints in an auxiliary vector . Applied to each row of it turns into the linear program (2.3), and it is also used to identify the cone in Proposition 3.
Formalization Note The statement is made for any whose polyhedron is nonempty, not only for the support of the §2 setting: nonemptiness is the only property of the proposition uses (strong LP duality for a feasible primal).
Preamble
import Mathlib import Definitions.Def_PrimalDualLDR_FixedRecourse_Basic import Definitions.Def_PrimalDualLDR_FixedRecourse_Setting import Definitions.Def_PrimalDualLDR_FixedRecourse_Problems open MeasureTheory Matrix
Formal statement
namespace PrimalDualLDR.FixedRecourse
/-- Proposition 1 (p. 5): for a nonempty polyhedron `Ξ = {ξ : Wξ ≥ h}` and any `z ∈ ℝ^k`,
`zᵀξ ≥ 0` for all `ξ ∈ Ξ` iff there is `λ ∈ ℝ^l` with `λ ≥ 0`, `Wᵀλ = z` and `hᵀλ ≥ 0`. -/
theorem proposition_1 {k l : ℕ} (W : Matrix (Fin l) (Fin k) ℝ) (h : Fin l → ℝ)
(hne : (polyhedron W h).Nonempty) (z : Fin k → ℝ) :
(∀ ξ ∈ polyhedron W h, 0 ≤ z ⬝ᵥ ξ) ↔
∃ lam : Fin l → ℝ, (∀ i, 0 ≤ lam i) ∧ Wᵀ.mulVec lam = z ∧ 0 ≤ h ⬝ᵥ lam := by sorry
end PrimalDualLDR.FixedRecourse
Source
Kuhn, Wiesemann, Georghiou, Primal and dual linear decision rules in stochastic and robust optimization, Optimization Online 2009/02/2218, p. 5, Proposition 1
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.