Proposition 7.7 — are positive homogeneous, subadditive and monotone
ProvedKallMayerSLP.RiskMeasures.proposition_7_7convex-analysisp2o-gran-per-chapterp2o-plan-bookp2o-strengthenedp2o-v1risk-measuresstochastic-programming
Let be a probability space and . For -dimensional random vectors with integrable components, let
Both are positive homogeneous and subadditive. Moreover is monotonically increasing and is monotonically decreasing with respect to the componentwise order.
These are the risk measures of joint integrated chance constraints (§4.1); by Proposition 7.1 they are convex.
Formalization Note The random vectors form (Lspace (Fin s → ℝ) 1 P), with components indexed from . The order hypothesis is componentwise and almost sure. The statement is made for every ; at both functionals vanish.
Differences from the source:
- reading: ϑ ∈ 𝓛¹₁ for the s-dimensional random vector ϑ (as printed in (P)) → ϑ ∈ 𝓛¹_s (Lspace (Fin s → ℝ) 1 P); reason: the printed subscript 1 is the dimension of a scalar variable; ϑ is s-dimensional, so the space is 𝓛¹_s of (1.6)
- reading: "monotonous" → ρ⁺_jic increasing, ρ⁻_jic decreasing, in the componentwise order; reason: the directions stated in the book's proof, p.189
- generalization: ϑ₁ ≤ ϑ₂ componentwise (pointwise) → ϑ₁ ≤ ϑ₂ almost surely; reason: a weaker hypothesis than the pointwise order of the book's proof ("ϑ₁(ω) ≤ ϑ₂(ω) for all ω"), so the monotonicity clause implies the book's
- encoding: 1 ≤ i ≤ s, s ≥ 1 implicit → i : Fin s (index base 0), any s ∈ ℕ; reason: at s = 0 both functionals are 0 and every clause holds trivially; for s ≥ 1 the same instances
Preamble
import Mathlib import Definitions.Def_KallMayerSLP_RiskMeasures_Lspace import Definitions.Def_KallMayerSLP_RiskMeasures_IsPosHomogeneous import Definitions.Def_KallMayerSLP_RiskMeasures_IsSubadditive import Definitions.Def_KallMayerSLP_RiskMeasures_rhoJicPlus import Definitions.Def_KallMayerSLP_RiskMeasures_rhoJicMinus open MeasureTheory open scoped ENNReal
Formal statement
namespace KallMayerSLP.RiskMeasures
/-- **Proposition 7.7** (joint integrated-chance-constraint risk measures).
Kall & Mayer, *Stochastic Linear Programming* (2005), Chapter 2, §7.2, (P) p.188 (PDF 198) and Prop. 7.7, p.189 (PDF 199), with the directions of
monotonicity from its proof. For an `s`-dimensional random vector `ϑ ∈ 𝓛¹_s`, both
`ρ⁺_jic(ϑ) := 𝔼[max_{1≤i≤s} ϑ_i⁺]` and `ρ⁻_jic(ϑ) := 𝔼[max_{1≤i≤s} ϑ_i⁻]` are positive
homogeneous and subadditive; moreover `ρ⁺_jic` is monotonically increasing and `ρ⁻_jic` is
monotonically decreasing.
**Formalization Note.** `𝓛¹_s` is `Lspace (Fin s → ℝ) 1 P` (integrable random vectors, index
base 0). The order on random vectors is componentwise, almost surely (`ϑ₁ ω i ≤ ϑ₂ ω i` for all `i`,
for `P`-almost every `ω`), weaker than pointwise, so the statement is stronger. Any `s` is allowed; for `s = 0` both functionals are `0`. -/
theorem proposition_7_7 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(s : ℕ) :
(IsPosHomogeneous (fun ϑ : Lspace (Fin s → ℝ) 1 P => rhoJicPlus P (ϑ : Ω → Fin s → ℝ)) ∧
IsSubadditive (fun ϑ : Lspace (Fin s → ℝ) 1 P => rhoJicPlus P (ϑ : Ω → Fin s → ℝ)) ∧
∀ ϑ₁ ∈ Lspace (Fin s → ℝ) 1 P, ∀ ϑ₂ ∈ Lspace (Fin s → ℝ) 1 P, ϑ₁ ≤ᵐ[P] ϑ₂ →
rhoJicPlus P ϑ₁ ≤ rhoJicPlus P ϑ₂) ∧
(IsPosHomogeneous (fun ϑ : Lspace (Fin s → ℝ) 1 P => rhoJicMinus P (ϑ : Ω → Fin s → ℝ)) ∧
IsSubadditive (fun ϑ : Lspace (Fin s → ℝ) 1 P => rhoJicMinus P (ϑ : Ω → Fin s → ℝ)) ∧
∀ ϑ₁ ∈ Lspace (Fin s → ℝ) 1 P, ∀ ϑ₂ ∈ Lspace (Fin s → ℝ) 1 P, ϑ₁ ≤ᵐ[P] ϑ₂ →
rhoJicMinus P ϑ₂ ≤ rhoJicMinus P ϑ₁) := by sorry
end KallMayerSLP.RiskMeasures
Source
Kall & Mayer, Stochastic Linear Programming (2005), Chapter 2, §7.2, (P), p.188 (PDF 198), and Proposition 7.7 with its proof, p.189 (PDF 199)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.