Eq. (39) — the transformed P-model feasible set is convex
ProvedChanceDetEquiv.PModel.convex_39chance-constrained-programmingconvex-optimizationfractional-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
For positive row and decision dimensions, P-model data and integrable moments, let contain all satisfying every constraint of program (39), including , , and all . Then
The claim includes points with , which are needed when comparing suprema of the two programs.
Formalization Note The fourth constraint uses with the raw second moment in (39b), not the centered variance. The paper's has , and the row residuals have Gaussian laws.
Preamble
import Definitions.Def_ChanceDetEquiv_PModel_Model
Formal statement
namespace ChanceDetEquiv.PModel
open MeasureTheory
/-- The feasible set of the transformed P-model program (39), including t = 0, is convex. -/
theorem convex_39 {Ω : Type*} [MeasurableSpace Ω] {m n : ℕ}
(d : Model Ω m n) [IsProbabilityMeasure d.P]
(hm : 0 < m) (hn : 0 < n)
(hMoments : MomentAssumptions d) (hNormal : NormalRows d)
(hConfidence : Confidence d) :
Convex ℝ (feasible39 d) := by sorry
end ChanceDetEquiv.PModel
Source
Charnes and Cooper, Deterministic Equivalents for Optimizing and Satisficing under Chance Constraints, Oper. Res. 11 (1963), pp. 32–33, Eqs. (39)–(39b) and paragraph after (39b); https://doi.org/10.1287/opre.11.1.18
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.