Proof of Theorem 1 — minimizers of lie in
ProvedWorstCaseCVaR.Mixture.two_convex_argminconvex-analysisp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be convex, and suppose the set of minimizers of is the nonempty closed bounded interval , . Let with . Then is convex and every minimizer of it lies in
This elementary fact about convex functions on the line localizes the minimizers of positive combinations of the component functions .
Formalization Note The paper calls the weights ; they are renamed to avoid a clash with the confidence level .
Preamble
import Mathlib import Definitions.Def_WorstCaseCVaR_Mixture_Setting open MeasureTheory
Formal statement
namespace WorstCaseCVaR.Mixture
theorem two_convex_argmin (g₁ g₂ : ℝ → ℝ)
(hg₁ : ConvexOn ℝ Set.univ g₁) (hg₂ : ConvexOn ℝ Set.univ g₂)
(a₁ b₁ a₂ b₂ : ℝ) (hab₁ : a₁ ≤ b₁) (hab₂ : a₂ ≤ b₂)
(hmin₁ : {t : ℝ | IsMinOn g₁ Set.univ t} = Set.Icc a₁ b₁)
(hmin₂ : {t : ℝ | IsMinOn g₂ Set.univ t} = Set.Icc a₂ b₂)
(c₁ c₂ : ℝ) (hc₁ : 0 ≤ c₁) (hc₂ : 0 ≤ c₂) (hc : 0 < c₁ + c₂) :
ConvexOn ℝ Set.univ (fun t => c₁ * g₁ t + c₂ * g₂ t) ∧
{t : ℝ | IsMinOn (fun t => c₁ * g₁ t + c₂ * g₂ t) Set.univ t} ⊆
Set.Icc (min a₁ a₂) (max b₁ b₂) := 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.