Proof of Theorem 1 — and the minimum is attained in
ProvedWorstCaseCVaR.Mixture.argmin_H_subsetconvex-analysiscvarmixture-distributionp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Under the standing assumptions ( probability distributions , integrable under each, ), suppose that for each the minimizer set of is , and let
Then for every :
- every minimizer of over lies in ;
- some point of minimizes over , so that .
This reduces the inner minimization to a compact interval independent of , which is what makes a minimax theorem applicable.
Preamble
import Mathlib import Definitions.Def_WorstCaseCVaR_Mixture_Setting open MeasureTheory
Formal statement
namespace WorstCaseCVaR.Mixture
theorem argmin_H_subset {l m n : ℕ} [NeZero l] (P : Fin l → Measure (Fin m → ℝ))
[∀ i, IsProbabilityMeasure (P i)] (f : (Fin n → ℝ) → (Fin m → ℝ) → ℝ)
(β : ℝ) (hβ0 : 0 < β) (hβ1 : β < 1) (x : Fin n → ℝ)
(hf : ∀ i, Integrable (f x) (P i))
(a b : Fin l → ℝ)
(hab : ∀ i, {α : ℝ | IsMinOn (Fi P f β x i) Set.univ α} = Set.Icc (a i) (b i)) :
∀ lam ∈ stdSimplex ℝ (Fin l),
{α : ℝ | IsMinOn (fun α => Hfun P f β x α lam) Set.univ α} ⊆
Set.Icc (Finset.univ.inf' Finset.univ_nonempty a)
(Finset.univ.sup' Finset.univ_nonempty b) ∧
∃ α₀ ∈ Set.Icc (Finset.univ.inf' Finset.univ_nonempty a)
(Finset.univ.sup' Finset.univ_nonempty b),
IsMinOn (fun α => Hfun P f β x α lam) Set.univ α₀ := by sorry
end WorstCaseCVaR.Mixture
Source
Zhu & Fukushima, Worst-Case Conditional Value-at-Risk with Application to Robust Portfolio Management, Oper. Res. 57(5), 2009, p. 1166, proof of Theorem 1
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.