§2.2, p. 5 — the a.s. constraints of hold on all of , and a.s. iff
ProvedPrimalDualLDR.FixedRecourse.sec_2_2_constraintsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1robust-optimizationstochastic-programming
Assume the standing assumptions of §2: is the support of , nonempty and bounded, of the form (2.1b), and spans . Let and . Then
- holds for -almost every if and only if it holds for every ;
- holds for -almost every if and only if it holds for every ;
- holds for -almost every if and only if
The first two items say that the almost-sure constraints of are robust constraints over the support ; the third replaces the semi-infinite equality constraint by a finite matrix equation. Together with Proposition 1 they turn into the linear program (2.3).
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
/-- §2.2, p. 5 (unnumbered): under the standing assumptions of §2, the almost-sure constraints of
`SP^u` hold for every `ξ` in the support `Ξ`, and the almost-sure equality constraint
`AXξ + Sξ = Bξ` is equivalent to the matrix equation `AX + S = B`. -/
theorem sec_2_2_constraints (σ : Setting) (hσ : σ.Standing)
(X : Matrix (Fin σ.n) (Fin σ.k) ℝ) (S : Matrix (Fin σ.m) (Fin σ.k) ℝ) :
((∀ᵐ ξ ∂σ.P, σ.A.mulVec (X.mulVec ξ) + S.mulVec ξ = σ.B.mulVec ξ) ↔
∀ ξ ∈ σ.Xi, σ.A.mulVec (X.mulVec ξ) + S.mulVec ξ = σ.B.mulVec ξ) ∧
((∀ᵐ ξ ∂σ.P, ∀ i, 0 ≤ (S.mulVec ξ) i) ↔ ∀ ξ ∈ σ.Xi, ∀ i, 0 ≤ (S.mulVec ξ) i) ∧
((∀ᵐ ξ ∂σ.P, σ.A.mulVec (X.mulVec ξ) + S.mulVec ξ = σ.B.mulVec ξ) ↔ σ.A * X + S = σ.B) := 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, §2.2, p. 5 (unnumbered claim)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.