Theorem 2, p. 910 (upper-bound part) — sup over 𝔽₂ of E_ℙ((r⁰ + r′ζ̃)⁺) ≤ π²(r⁰, r)
ProvedTractableDRO.MeanCov.theorem_2_upperdistributionally-robust-optimizationmoment-boundsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let , , , and . Let be the family of probability distributions of a random vector with finite second moments, mean and . Then
Both sides are extended reals: the left side is when is empty and the right side is when has no solution.
This is the bound half of Theorem 2 of Goh and Sim: the deterministic problem is a valid, tractable upper bound on the worst-case expected positive part over all distributions with the prescribed mean set and covariance.
Formalization Note No hypothesis on , or is needed for this half: membership in already forces , which is positive semidefinite. Objects are those of TractableDRO.MeanCov.Model.
Preamble
import Mathlib import Definitions.Def_TractableDRO_MeanCov_Model open Matrix
Formal statement
namespace TractableDRO.MeanCov
theorem theorem_2_upper {n N : ℕ} (F : Matrix (Fin N) (Fin n) ℝ)
(Sig : Matrix (Fin N) (Fin N) ℝ) (Vhat : Set (Fin n → ℝ)) (r0 : ℝ) (r : Fin n → ℝ) :
TractableDRO.MeanSupport.worstCase (family2 F Sig Vhat) r0 r ≤ pi2 F Sig Vhat r0 r := by sorry
end TractableDRO.MeanCov
Source
Goh & Sim, Distributionally Robust Optimization and Its Tractable Approximations, Oper. Res. 58(4), 2010, p. 910, Theorem 2 and (22)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.