Convexity, §2, p. 3 — subadditivity and positive homogeneity imply convexity of ρ
ProvedCoherentSDDP.Inner.convexity_of_coherentcoherent-risk-measureconvexityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be a set and let map real random variables to . Suppose is subadditive, , and positively homogeneous, for . Then is convex: for all and every ,
In the paper, convexity together with monotonicity is what keeps the nested problems (1)–(2) convex, and it is the property of used in the base case of the proof of Proposition 5.
Formalization Note Only the two axioms named on the page are assumed. At both sides coincide, so positive homogeneity is not needed at .
Preamble
import Mathlib import Definitions.Def_CoherentSDDP_Inner_Basic
Formal statement
namespace CoherentSDDP.Inner
theorem convexity_of_coherent {Ω : Type*} (ρ : (Ω → ℝ) → ℝ) (hsub : RiskSubadditive ρ)
(hhom : RiskPosHomogeneous ρ) :
∀ (Z₁ Z₂ : Ω → ℝ) (α : ℝ), 0 ≤ α → α ≤ 1 →
ρ (α • Z₁ + (1 - α) • Z₂) ≤ α * ρ Z₁ + (1 - α) * ρ Z₂ := by sorry
end CoherentSDDP.Inner
Source
Philpott, de Matos & Finardi, On Solving Multistage Stochastic Programs with Coherent Risk Measures, authors' manuscript of 13 August 2012 (Operations Research, 2013), p. 3, §2, Convexity
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.