(RC[]) is less conservative than (IRC[])
ProvedRobustLP.Counterpart.irc_feasible_rc_feasiblelinear-programmingp2o-batch-p200bp2o-gran-per-chapterp2o-plan-paperp2o-v1robust-optimization
Let an uncertain linear program with uncertain-entry sets be given, and let , , . If is feasible for the interval robust counterpart (IRC[ε, δ]), then with
is feasible for the robust counterpart (RC[ε, δ, Ω]).
Every admissible for the linear counterpart is therefore admissible for the conic one, so the optimal value of (RC) is at most that of (IRC).
Formalization Note The page does not state the sign of ; the hypothesis is carried for uniformity with Proposition 1, where (RC) is introduced with a positive parameter.
Preamble
import Mathlib import Definitions.Def_RobustLP_Counterpart_UncertainLP import Definitions.Def_RobustLP_Counterpart_RobustCounterparts
Formal statement
namespace RobustLP.Counterpart
/-- **(RC) is less conservative than (IRC)** (Ben-Tal–Nemirovski 2000, §3.1, p. 420). For
`ε > 0`, `δ > 0`, `Ω > 0`: if `(x, y)` is feasible for (IRC[ε, δ]), then `(x, y', z')` with
`y'_{ij} = y_j` and `z'_{ij} = 0` is feasible for (RC[ε, δ, Ω]). -/
theorem irc_feasible_rc_feasible {n p m : ℕ} (L : UncertainLP n p m)
(ε δ Ω : ℝ) (hε : 0 < ε) (hδ : 0 < δ) (hΩ : 0 < Ω) (x y : Fin n → ℝ)
(hIRC : L.IRCFeasible ε δ x y) :
L.RCFeasible ε δ Ω x (fun _ j => y j) (fun _ _ => 0) := by sorry
end RobustLP.Counterpart
Source
Ben-Tal and Nemirovski, Robust solutions of Linear Programming problems contaminated with uncertain data, Math. Program. Ser. A 88 (2000) 411–424, p. 420, §3.1, 'Note that (RC[ϵ, δ, Ω]) is "less conservative" that (IRC[ϵ, δ])'
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.