Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 4, p. 17 — for r > 0 the relative-entropy DRO predictor ĉ_r is strongly optimal in (5)

Proved
DROOptimal.Predictor.theorem_4

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

distributionally-robust-optimizationlarge-deviationsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1relative-entropyvector-optimization

Let X⊆RnX\subseteq\mathbb R^nX⊆Rn be compact, Ξ={1,…,d}\Xi=\{1,\dots,d\}Ξ={1,…,d}, and γ:X×Ξ→R\gamma:X\times\Xi\to\mathbb Rγ:X×Ξ→R continuous in xxx for each iii; write c(x,P)=∑iP(i)γ(x,i)c(x,\mathbb P)=\sum_i\mathbb P(i)\gamma(x,i)c(x,P)=∑i​P(i)γ(x,i) for the expected cost under a model P\mathbb PP in the simplex P\mathcal PP, III for the relative entropy, and P^T\hat{\mathbb P}_TP^T​ for the empirical distribution of TTT independent samples. For r>0r>0r>0 let

c^r(x,P′)=sup⁡P∈P{c(x,P):I(P′,P)≤r}.\hat c_r(x,\mathbb P')=\sup_{\mathbb P\in\mathcal P}\{c(x,\mathbb P): I(\mathbb P',\mathbb P)\le r\}.c^r​(x,P′)=P∈Psup​{c(x,P):I(P′,P)≤r}.

Then c^r\hat c_rc^r​ is strongly optimal in the vector optimization problem (5):

  1. c^r\hat c_rc^r​ is continuous on X×PX\times\mathcal PX×P, and for all x∈Xx\in Xx∈X and P∈P\mathbb P\in\mathcal PP∈P,
lim sup⁡T→∞1Tlog⁡P∞(c(x,P)>c^r(x,P^T))≤−r;\limsup_{T\to\infty}\frac1T\log\mathbb P^\infty\big(c(x,\mathbb P)>\hat c_r(x,\hat{\mathbb P}_T)\big)\le-r ;T→∞limsup​T1​logP∞(c(x,P)>c^r​(x,P^T​))≤−r;
  1. every continuous c^:X×P→R\hat c:X\times\mathcal P\to\mathbb Rc^:X×P→R with the same property satisfies c^r(x,P′)≤c^(x,P′)\hat c_r(x,\mathbb P')\le\hat c(x,\mathbb P')c^r​(x,P′)≤c^(x,P′) for all x∈Xx\in Xx∈X and P′∈P\mathbb P'\in\mathcal PP′∈P.

In words: among all data-driven predictors whose out-of-sample disappointment decays at exponential rate at least rrr under every model, the relative-entropy distributionally robust predictor is the least conservative, uniformly in the decision and the observed frequencies.

Formalization Note The rate condition is the logarithm-free RateLE (equivalent to the displayed lim sup⁡\limsuplimsup with log⁡0=−∞\log0=-\inftylog0=−∞); the competitors range over all jointly continuous functions on X×PX\times\mathcal PX×P and nothing else; feasibility of c^r\hat c_rc^r​, continuity included, is part of the conclusion.

Preamble
import Mathlib
import Definitions.Def_DROOptimal_Predictor_Setting
import Definitions.Def_DROOptimal_Predictor_Problem5
Formal statement
namespace DROOptimal.Predictor

/-- Theorem 4 (Optimality of ĉ_r), p. 17: under the standing assumptions of §2 (p. 5), if r > 0
then ĉ_r is strongly optimal in (5): it is a continuous predictor whose disappointment decays at rate
at least r under every model and decision, and ĉ_r(x, ℙ′) ≤ ĉ(x, ℙ′) for all x ∈ X, ℙ′ ∈ 𝒫 and every
predictor ĉ feasible in (5). -/
theorem theorem_4 {n d : ℕ} (X : Set (EuclideanSpace ℝ (Fin n))) (hX : IsCompact X)
    (γ : ↥X → Fin d → ℝ) (hγ : ∀ i, Continuous (fun x : ↥X => γ x i))
    (r : ℝ) (hr : 0 < r) :
    StronglyOptimal5 γ r (drPredictor γ r) := by sorry

end DROOptimal.Predictor
Source
Van Parys, Mohajerin Esfahani & Kuhn, From Data to Decisions: Distributionally Robust Optimization is Optimal, arXiv:1704.04118v3, p. 17, Theorem 4; proof pp. 17–19
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 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