(17), Sect. 5.1.1, p. 14 — ℙᴺ{V(x*_N) > ε(s*_N)} = Σ_k C(N,k) ∫_(ε(k),1] (1−v)^(N−k) dF_k(v)
ProvedWaitJudge.Convex.prob_violation_eq_sum_integralp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1scenario-optimizationsupport-constraintswait-and-judge
Consider a convex scenario program: , every and the tie-break functions are convex, the sample is i.i.d. with law , , and Assumptions 1 and 2 and the measurability conditions hold. Let for . Then
where is the generalized distribution function (12).
This is the formula by which the probability that the violation exceeds the a-posteriori level is computed from alone; Theorem 1 is obtained by maximising its right-hand side over all compatible with convex problems.
Formalization Note is a finite measure on , and the probability is compared with the real right-hand side through .
Preamble
import Mathlib import Definitions.Def_ScenarioApproach_Generalization_violation import Definitions.Def_WaitJudge_Convex_Setting open MeasureTheory ScenarioApproach.Generalization
Formal statement
namespace WaitJudge.Convex
theorem prob_violation_eq_sum_integral {d p : ℕ} {Δ : Type*} [MeasurableSpace Δ] (P : Measure Δ) [IsProbabilityMeasure P]
(c : E d) (X : Set (E d)) (Xδ : Δ → Set (E d)) (tb : Fin p → E d → ℝ)
(hX : Convex ℝ X) (hXδ : ∀ δ, Convex ℝ (Xδ δ))
(htb : ∀ j, ConvexOn ℝ Set.univ (tb j))
(hA1 : Assumption1 c X Xδ tb) (hA2 : Assumption2 c X Xδ tb P)
(hmeas : MeasurabilityPins c X Xδ tb)
(N : ℕ) (hN : d < N) (ε : ℕ → ℝ) (hε : ∀ k ≤ d, 0 ≤ ε k ∧ ε k ≤ 1) :
(Measure.pi fun _ : Fin N => P) {ω | ε (sstar c X Xδ tb ω) < violation P Xδ (xstar c X Xδ tb ω)} =
ENNReal.ofReal (∑ k ∈ Finset.range (d + 1),
(N.choose k : ℝ) * ∫ v in Set.Ioc (ε k) 1, (1 - v) ^ (N - k) ∂(Fk P c X Xδ tb k)) := by sorry
end WaitJudge.Convex
Source
Campi & Garatti, Wait-and-judge scenario optimization, Math. Program. (2018), doi:10.1007/s10107-016-1056-9 (accepted manuscript), PDF p. 14, Sect. 5.1.1, (17) (from (13), (14), (16))
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.