Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 6.2, p. 20 — Algorithm 1 offers sub-optimal assortments in at most Nτ good epochs

Proved
MNLBandit.WellSeparated.lemma_6_2

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

assortment-optimizationinstance-dependent-regretmnl-banditp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Consider the MNL-Bandit problem under Assumption 4.1 (0≤vi≤v0=10\le v_i\le v_0=10≤vi​≤v0​=1, a feasible family S\mathcal SS of the form (2.3) closed under subsets), with revenues ri∈[0,1]r_i\in[0,1]ri​∈[0,1], and run Algorithm 1 with any tie-breaking rule along any history of TTT customers. Let τ=4NClog⁡NT/Δ2(v)\tau=4NC\log NT/\Delta^2(\mathbf v)τ=4NClogNT/Δ2(v) as in (6.2), and let LLL be the number of epochs started within the horizon. Then

∣{ℓ∈{0,…,L−1}:epoch ℓ is good and R(Sℓ+1,v)<R(S∗,v)}∣≤Nτ.\Big|\big\{\ell\in\{0,\dots,L-1\} : \text{epoch }\ell\text{ is good and }R(S_{\ell+1},\mathbf v)<R(S^*,\mathbf v)\big\}\Big|\le N\tau.​{ℓ∈{0,…,L−1}:epoch ℓ is good and R(Sℓ+1​,v)<R(S∗,v)}​≤Nτ.

In words: Algorithm 1 cannot offer sub-optimal assortments in more than NτN\tauNτ good epochs.

Together with the fact that bad epochs are rare, this bounds the number of epochs in which the algorithm loses revenue. This is the step that gives the logarithmic dependence on TTT in Theorem 3.

Formalization Note As in Lemma 6.1, a good epoch ℓ\ellℓ is paired with the assortment Sℓ+1S_{\ell+1}Sℓ+1​ selected by its upper confidence bounds. Epoch ℓ=0\ell=0ℓ=0 (the first assortment, chosen from vi,0UCB=1v^{\mathrm{UCB}}_{i,0}=1vi,0UCB​=1) is included in the count. The bound is pathwise: it holds along every history.

Preamble
import Mathlib
import Definitions.Def_ChoiceCDLP_MNL_mnlObjective
import Definitions.Def_MNLBandit_WellSeparated_Setting
Formal statement
namespace MNLBandit.WellSeparated

open ChoiceCDLP.MNL

open Classical in
/-- Lemma 6.2, p. 20, for Algorithm 1 under Assumption 4.1 and `r_i ∈ [0, 1]`, along any history
`h` of `T` customers: among the epochs started within the horizon, at most `Nτ` are good epochs
offering a sub-optimal assortment (epoch `ℓ + 1` counted when the UCBs computed at the end of
epoch `ℓ` are good and `R(S_{ℓ+1}, v) < R(S*, v)`). -/
theorem lemma_6_2 (N : ℕ) (v r : Fin N → ℝ) (𝒮 : Finset (Finset (Fin N))) (h𝒮 : 𝒮.Nonempty)
    (sel : (Finset (Fin N) → ℝ) → Finset (Fin N))
    (hTU : MNLBandit.UCB.IsTU 𝒮) (hdown : MNLBandit.UCB.DownClosed 𝒮) (hv : ∀ i, 0 ≤ v i ∧ v i ≤ 1)
    (hr : ∀ i, r i ∈ Set.Icc (0 : ℝ) 1) (hsel : MNLBandit.UCB.IsArgmaxSel 𝒮 sel)
    (T : ℕ) (h : Fin T → Option (Fin N)) :
    (((Finset.range (epochsStarted r sel h)).filter (fun ℓ =>
        IsGood v r sel h ℓ ∧
          revenue v r (epochSet r sel h (ℓ + 1)) < MNLBandit.UCB.optRevenue v r 𝒮 h𝒮)).card : ℝ)
      ≤ N * tau N T (gap v r 𝒮 h𝒮) := by sorry

end MNLBandit.WellSeparated
Source
Agrawal, Avadhanula, Goyal, Zeevi, MNL-Bandit: A Dynamic Learning Approach to Assortment Selection, arXiv:1706.03880v2, p. 20, Lemma 6.2 (proof pp. 46–47, (C.2)–(C.4))
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