The globalized robust counterpart: perturbation structures and the consistent distance
DefinitionRobustOptimization_GRCconvex-optimizationlinear-optimizationoptimization
Retired 2026-09-21. This declaration belongs to the first draft of the Ben-Tal, El Ghaoui and Nemirovski, Robust Optimization (2009) mission series, which was withdrawn in full. The series is being redrafted from the source under the namespace RobustOpt; build on those declarations, not on this one. The original statement and prose are preserved in the Prove2OR repository, branch textbook-sequential, under papers/ben-tal-2009-robust-optimization/retired/.
Definition code
import Definitions.Def_RobustOptimization_LO
import Mathlib
namespace RobustOptimization
open Matrix Finset
/-- The distance `dist(ζ, Z | L) = inf {‖ζ - ζ'‖ : ζ' ∈ Z, ζ - ζ' ∈ L}` of (3.1.3), measuring
how far a physically possible perturbation is from the normal range, consistently with the
decomposition `Z₊ = Z + L`. -/
noncomputable def DistToNormalRange {L : ℕ} (Z : Set (Fin L → ℝ)) (C : Set (Fin L → ℝ))
(nrm : (Fin L → ℝ) → ℝ) (ζ : Fin L → ℝ) : ℝ :=
sInf {t : ℝ | ∃ ζ' ∈ Z, ζ - ζ' ∈ C ∧ t = nrm (ζ - ζ')}
/-- `x` is a globally robust feasible solution with global sensitivity `α` (Definition 3.1.1):
the uncertain linear constraint may be violated on the physically possible perturbations
`Z₊ = Z + L`, but only in proportion to the distance to the normal range `Z`. -/
def GloballyRobustFeasible {n L : ℕ} (a0 : Fin n → ℝ) (a : Fin L → Fin n → ℝ) (b0 : ℝ)
(b : Fin L → ℝ) (Z : Set (Fin L → ℝ)) (C : Set (Fin L → ℝ))
(nrm : (Fin L → ℝ) → ℝ) (α : ℝ) (x : Fin n → ℝ) : Prop :=
∀ ζ' ∈ Z, ∀ δ ∈ C, (a0 + ∑ l, (ζ' + δ) l • a l) ⬝ᵥ x
≤ b0 + (∑ l, (ζ' + δ) l * b l) + α * DistToNormalRange Z C nrm (ζ' + δ)
/-- A perturbation structure `(Z, L, ‖·‖)` of (G.a)-(G.c): a nonempty closed convex normal
range, a closed convex cone of physically possible directions, and a norm on the perturbation
space. -/
def IsPerturbationStructure {L : ℕ} (Z : Set (Fin L → ℝ)) (C : Set (Fin L → ℝ))
(nrm : (Fin L → ℝ) → ℝ) : Prop :=
(Z.Nonempty ∧ IsClosed Z ∧ Convex ℝ Z) ∧
(IsClosed C ∧ Convex ℝ C ∧ ∀ c : ℝ, 0 ≤ c → ∀ z ∈ C, c • z ∈ C) ∧
((∀ v, 0 ≤ nrm v) ∧ (∀ v, nrm v = 0 ↔ v = 0) ∧
(∀ (c : ℝ) (v), nrm (c • v) = |c| * nrm v) ∧ ∀ v w, nrm (v + w) ≤ nrm v + nrm w)
end RobustOptimizationSource
Aharon Ben-Tal, Laurent El Ghaoui and Arkadi Nemirovski, Robust Optimization, Princeton Series in Applied Mathematics, Princeton University Press, 2009, https://doi.org/10.1515/9781400831050, p. 68: the conditions (G.a)-(G.c), the sum set (3.1.2), the distance (3.1.3), and Definition 3.1.1 with the semi-infinite constraint (3.1.4). Source PDF sha256 b4aba60d1c863237fe4ee6ce4b62051d945cddbcb78523d03b06d6eff13a7c91.