Theorem 1, p. 909 (upper-bound part) — sup over 𝔽₁ of E_ℙ((r⁰ + r′ζ̃)⁺) ≤ π¹(r⁰, r)
ProvedTractableDRO.MeanSupport.theorem_1_upperdistributionally-robust-optimizationdualitymoment-boundsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
This is the upper-bound half of Theorem 1 of Goh and Sim (2010).
Let be arbitrary sets, and let be the family of probability distributions of a random vector with integrable coordinates, mean and . For every and ,
both sides taken in the extended reals.
The bound is a deterministic convex program over the sets and , so this inequality turns an optimization over distributions into a tractable robust optimization problem.
Formalization Note No convexity, closedness or nonemptiness is assumed: the inequality holds for all sets. When is empty the left side is (⊥). All values are in EReal; see the definition TractableDRO.MeanSupport.Model.
Preamble
import Mathlib import Definitions.Def_TractableDRO_MeanSupport_Model open MeasureTheory
Formal statement
namespace TractableDRO.MeanSupport
theorem theorem_1_upper {n : ℕ} (V Vhat : Set (Fin n → ℝ)) (r0 : ℝ) (r : Fin n → ℝ) :
worstCase (family1 V Vhat) r0 r ≤ pi1 V Vhat r0 r := by sorry
end TractableDRO.MeanSupport
Source
Goh & Sim, Distributionally Robust Optimization and Its Tractable Approximations, Oper. Res. 58(4), 2010, p. 909, Theorem 1 (upper-bound part), (21)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.