Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proof of Theorem 2.2, p. 586 — Jensen: EP[Q(r~,x)]≤Q(EP[r~],x)\mathbb E_P[\mathcal Q(\tilde r,x)]\le\mathcal Q(\mathbb E_P[\tilde r],x)EP​[Q(r~,x)]≤Q(EP​[r~],x), and the expectation is finite

Proved
MinimaxSLP.ObjExtremal.jensen_step

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

jensen-inequalityp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1second-stage-value

Let xxx be such that X(x)≠∅X(x)\neq\emptysetX(x)=∅, and assume Assumption 3. Let PPP be a probability distribution of a random vector r~\tilde rr~ on Rd\mathbb R^dRd whose coordinates are integrable. Then Q(r~,x)\mathcal Q(\tilde r,x)Q(r~,x) is integrable under PPP, in particular EP[Q(r~,x)]>−∞\mathbb E_P[\mathcal Q(\tilde r,x)]>-\inftyEP​[Q(r~,x)]>−∞, and

EP[Q(r~,x)] ≤ Q(EP[r~],x).\mathbb E_P\big[\mathcal Q(\tilde r,x)\big]\ \le\ \mathcal Q\big(\mathbb E_P[\tilde r],x\big).EP​[Q(r~,x)] ≤ Q(EP​[r~],x).

Moreover Q(0,x)=0\mathcal Q(0,x)=0Q(0,x)=0, so the right-hand side vanishes when EP[r~]=0\mathbb E_P[\tilde r]=0EP​[r~]=0.

In the proof of Theorem 2.2 this is applied to the centred normal vectors r~k\tilde r_kr~k​, and shows that the ϵ\sqrt\epsilonϵ​ correction term is bounded below and vanishes as ϵ↓0\epsilon\downarrow0ϵ↓0.

Formalization Note The paper applies the inequality to r~k∼N(0,(Vkvk0−vkvk′)/vk02)\tilde r_k\sim\mathbb N\big(0,(V_kv_{k0}-v_kv_k')/v_{k0}^2\big)r~k​∼N(0,(Vk​vk0​−vk​vk′​)/vk02​); the statement here is the general form for any probability measure with integrable coordinates. The paper attributes finiteness to "Assumptions 1–4"; only X(x)≠∅X(x)\neq\emptysetX(x)=∅ and Assumption 3 are needed.

Preamble
import Mathlib
import Definitions.Def_MomentDRO_Conf_Setting
import Definitions.Def_MinimaxSLP_ObjExtremal_Model

open MeasureTheory Matrix Filter Topology
Formal statement
namespace MinimaxSLP.ObjExtremal

/-- Proof of Theorem 2.2 (p. 586), Jensen step, in general form: for every probability measure
`P` on `ℝ^d` with integrable coordinates, `𝒬(·, x)` is `P`-integrable (so its expectation is
`> −∞`) and `E_P[𝒬(r̃, x)] ≤ 𝒬(E_P[r̃], x)`; moreover `𝒬(0, x) = 0`, the value at mean `0`. -/
theorem jensen_step {n r d : ℕ}
    (W : Matrix (Fin r) (Fin d) ℝ) (T : Matrix (Fin r) (Fin n) ℝ) (h : Fin r → ℝ)
    (hA3 : ∀ q : Fin d → ℝ, ∃ π : Fin r → ℝ, Wᵀ *ᵥ π ≤ q)
    (x : Fin n → ℝ) (hrec : (MinimaxSLP.ObjSDP.recourseSet W T h x).Nonempty)
    (P : Measure (Fin d → ℝ)) [IsProbabilityMeasure P]
    (hint : ∀ i : Fin d, Integrable (fun ρ : Fin d → ℝ => ρ i) P) :
    Integrable (fun ρ => MinimaxSLP.ObjSDP.Qval W T h ρ x) P ∧
    ∫ ρ, MinimaxSLP.ObjSDP.Qval W T h ρ x ∂P ≤ MinimaxSLP.ObjSDP.Qval W T h (fun i => ∫ ρ, ρ i ∂P) x ∧
    MinimaxSLP.ObjSDP.Qval W T h 0 x = 0 := by sorry

end MinimaxSLP.ObjExtremal
Source
Bertsimas, Doan, Natarajan & Teo, Models for Minimax Stochastic Linear Optimization Problems with Risk Aversion, Math. Oper. Res. 35(3), 2010, p. 586, proof of Theorem 2.2
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me