Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 5.1, p. 15 — every online algorithm has expected regret at least ϵT/6 on I_MAB

Proved
MNLBandit.LowerBound.lemma_5_1

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

lower-boundmulti-armed-banditp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1regret

Let N≥2N\ge2N≥2 arms, a horizon T≥1T\ge1T≥1, α>0\alpha>0α>0, and

ϵ=1100NαT,\epsilon=\frac1{100}\sqrt{\frac{N\alpha}{T}},ϵ=1001​TNα​​,

with 2α+ϵ≤12\alpha+\epsilon\le12α+ϵ≤1. On the randomized Bernoulli instance IMABI_{\mathrm{MAB}}IMAB​ (Definition 5.1: a hidden arm jjj, uniform on {1,…,N}\{1,\dots,N\}{1,…,N}, has mean α+ϵ\alpha+\epsilonα+ϵ and every other arm mean α\alphaα), every online algorithm A\mathcal AA, deterministic or randomized, that plays arm At\mathcal A_tAt​ at time ttt satisfies

RegA(T,μ)=E[∑t=1T(μj−μAt)]≥ϵT6,\mathrm{Reg}_{\mathcal A}(T,\mu)=\mathbb E\Big[\sum_{t=1}^{T}(\mu_j-\mu_{\mathcal A_t})\Big]\ge\frac{\epsilon T}{6},RegA​(T,μ)=E[t=1∑T​(μj​−μAt​​)]≥6ϵT​,

where the expectation is over the hidden arm jjj and over the rewards of the pulled arms (and the algorithm's own randomization).

This MAB lower bound is what the reduction of §5.2 transfers to the MNL-Bandit problem under a cardinality constraint.

Formalization Note The page assumes only α<1\alpha<1α<1. That is not enough: for α\alphaα close to 111 the biased arm is identified after few pulls and the bound fails; the proof sets "α\alphaα to be a small enough constant" (p. 19). The added hypothesis 2α+ϵ≤12\alpha+\epsilon\le12α+ϵ≤1 (which implies α<1\alpha<1α<1) and α>0\alpha>0α>0 restore it. T≥1T\ge1T≥1 is added because ϵ\epsilonϵ has TTT in a denominator. The printed proof's line "Let ϵ=N/(60αT)\epsilon=\sqrt{N/(60\alpha T)}ϵ=N/(60αT)​" contradicts Definition 5.1; the lemma is stated with Definition 5.1's ϵ\epsilonϵ, which satisfies the proof's requirement T≤Nα/(60ϵ2)T\le N\alpha/(60\epsilon^2)T≤Nα/(60ϵ2). The printed N≥2N\ge2N≥2 is kept although the printed proof (via Lemma E.2) needs N≥12N\ge12N≥12.

Preamble
import Mathlib
import Definitions.Def_MNLBandit_LowerBound_BernoulliMAB
Formal statement
namespace MNLBandit.LowerBound

theorem lemma_5_1 (N : ℕ) (hN : 2 ≤ N) (α : ℝ) (hα : 0 < α) (T : ℕ) (hT : 1 ≤ T)
    (hαϵ : 2 * α + epsMAB N α T ≤ 1) (A : MABAlg N) (hA : IsMABAlg A) :
    epsMAB N α T * T / 6 ≤ regretIMAB A α (epsMAB N α T) T := by sorry

end MNLBandit.LowerBound
Source
Agrawal, Avadhanula, Goyal, Zeevi, MNL-Bandit: A Dynamic Learning Approach to Assortment Selection, arXiv:1706.03880v2, p. 15, Lemma 5.1 and Definition 5.1; proof p. 58
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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