Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 4 — χ² and G uncertainty sets bound worst-case VaR

Proved
DataDrivenRO.Discrete.theorem_4

by mikedeng1 · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

p2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1robust-optimizationstatisticsuncertainty-setvalue-at-risk

Fix listed support vectors aj∈Rda_j\in\mathbb R^daj​∈Rd, an empirical probability vector p^∈Δn\hat p\in\Delta_np^​∈Δn​, and a nonnegative test radius ρ\rhoρ. The Pearson and G confidence regions of (10) determine the uncertainty sets of (12) and (13). Simultaneously for every tail level 0<ϵ<10<\epsilon<10<ϵ<1, every law PpP_pPp​ with ppp in either region, and every direction vvv,

VaR⁡ϵPp(v)≤δ∗(v∣Uϵχ2)(p∈Pχ2),VaR⁡ϵPp(v)≤δ∗(v∣UϵG)(p∈PG).\operatorname{VaR}^{P_p}_{\epsilon}(v) \le\delta^*(v\mid U^{\chi^2}_\epsilon) \quad(p\in\mathcal P^{\chi^2}),\qquad \operatorname{VaR}^{P_p}_{\epsilon}(v) \le\delta^*(v\mid U^G_\epsilon) \quad(p\in\mathcal P^G).VaRϵPp​​(v)≤δ∗(v∣Uϵχ2​)(p∈Pχ2),VaRϵPp​​(v)≤δ∗(v∣UϵG​)(p∈PG).

Both uncertainty sets are nonempty, convex, and compact. Thus their support functions are genuine finite suprema, and Theorem 1's geometric criterion turns these bounds into probabilistic guarantees whenever the confidence regions cover the true law.

Formalization Note This statement formalizes Theorem 4's deterministic risk-bound content. The paper's sampling probability comes from coverage of the Pearson and G tests, which is not asserted here. The test threshold is represented by an arbitrary ρ≥0\rho\ge0ρ≥0, corresponding to χn−1,1−α2/(2N)\chi^2_{n-1,1-\alpha}/(2N)χn−1,1−α2​/(2N). The explicit conic support-function programs (14)–(15) are separate, unformalized companions. The set of listed support points is represented by Fin n, and PpP_pPp​ is the corresponding point-mass law.


Retired 2026-10-27: a Lean-vs-paper fidelity audit found this statement weaker than, or narrower than, the cited result. A corrected statement will be resubmitted.

Preamble
import Mathlib
import Definitions.Def_DataDrivenRO_Discrete_Setting
Formal statement
namespace DataDrivenRO.Discrete

/-- Deterministic content of Theorem 4, p. 13: for every tail level, both uncertainty sets
bound the Value at Risk of each law in the corresponding confidence region.  Their geometric
properties make the real support functions genuine and allow Theorem 1(a) to apply. -/
theorem theorem_4 {d n : ℕ} (a : Fin n → (Fin d → ℝ))
    (phat : Fin n → ℝ) (hphat : phat ∈ stdSimplex ℝ (Fin n))
    (ρ : ℝ) (hρ : 0 ≤ ρ) :
    ∀ ε : ℝ, 0 < ε → ε < 1 →
      (∀ p ∈ chi2Region phat ρ, ∀ v : Fin d → ℝ,
        VaR a p ε v ≤ RobustMDP.Shared.supportFunction (Uchi2 a phat ρ ε) v) ∧
      (∀ p ∈ gRegion phat ρ, ∀ v : Fin d → ℝ,
        VaR a p ε v ≤ RobustMDP.Shared.supportFunction (UG a phat ρ ε) v) ∧
      (Uchi2 a phat ρ ε).Nonempty ∧ Convex ℝ (Uchi2 a phat ρ ε) ∧
        IsCompact (Uchi2 a phat ρ ε) ∧
      (UG a phat ρ ε).Nonempty ∧ Convex ℝ (UG a phat ρ ε) ∧
        IsCompact (UG a phat ρ ε) := by sorry

end DataDrivenRO.Discrete
Source
Bertsimas, Gupta & Kallus, Data-Driven Robust Optimization, arXiv:1401.0212v2, Theorem 4 and (12)–(13), p. 13; proof, p. ec2
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 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