Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma D.1, PDF pp. 46–47 — shrinking MNL attraction values does not raise the optimal revenue

Proved
MultiStoreAssort.Approx.lemmaD1_attraction_monotone

by mikedeng1 · Oct 14, 2026 · Mathlib 0df444a (Lean v4.33.1)

assortment-optimizationmultinomial-logitp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Consider an MNL model with nnn products, revenues r1,…,rn∈Rr_1,\dots,r_n\in\mathbb Rr1​,…,rn​∈R and attraction values v1,…,vn>0v_1,\dots,v_n>0v1​,…,vn​>0; the outside option has attraction value 111. Let

R=max⁡S∑i∈Srivi1+∑i∈Svi,R=\max_{S}\frac{\sum_{i\in S}r_iv_i}{1+\sum_{i\in S}v_i},R=Smax​1+∑i∈S​vi​∑i∈S​ri​vi​​,

the maximum taken over all assortments S⊆{1,…,n}S\subseteq\{1,\dots,n\}S⊆{1,…,n}. Given ϵ1,…,ϵn\epsilon_1,\dots,\epsilon_nϵ1​,…,ϵn​ with 0<ϵi<10<\epsilon_i<10<ϵi​<1, let

R′=max⁡S∑i∈Sriϵivi1+∑i∈Sϵivi.R'=\max_{S}\frac{\sum_{i\in S}r_i\epsilon_iv_i}{1+\sum_{i\in S}\epsilon_iv_i}.R′=Smax​1+∑i∈S​ϵi​vi​∑i∈S​ri​ϵi​vi​​.

Then R≥R′R\ge R'R≥R′.

The lemma says that the optimal MNL revenue is monotone in the attraction values. In the proof of Proposition 7 it bounds the revenue a store earns from the products of other stores, whose attraction values are discounted by exp⁡(−cik(S))\exp(-c^k_i(\boldsymbol S))exp(−cik​(S)), by the single-store optimum.

Formalization Note. The two optimal values are not formed as terms; the statement is: for every assortment TTT there is an assortment SSS whose revenue under vvv is at least the revenue of TTT under ϵv\epsilon vϵv. No order or sign is assumed on the revenues, as in the lemma.

Preamble
import Mathlib
import Definitions.Def_ChoiceCDLP_MNL_mnlObjective
import Definitions.Def_MultiStoreAssort_Approx_Model
Formal statement
namespace MultiStoreAssort.Approx

/-- Lemma D.1 (Online Supplement pp. 14–15): scaling every attraction value down by a factor
`ε_i ∈ (0, 1)` does not increase the optimal MNL revenue: `R ≥ R'`, where
`R = max_S MultiStoreAssort.SMR.single r v S` and `R' = max_S MultiStoreAssort.SMR.single r (ε v) S`. -/
theorem lemmaD1_attraction_monotone {n : ℕ} (r v ε : Fin n → ℝ) (hv : ∀ i, 0 < v i)
    (hε : ∀ i, 0 < ε i ∧ ε i < 1) :
    ∀ T : Finset (Fin n), ∃ S : Finset (Fin n), MultiStoreAssort.SMR.single r (fun i => ε i * v i) T ≤ MultiStoreAssort.SMR.single r v S := by sorry

end MultiStoreAssort.Approx
Source
Liu, Chen, Liu & Wan, Simultaneous vs. Sequential: Optimal Assortment Recommendation in Multistore Retailing, authors' manuscript with Online and Additional Supplements (2024-09-17; MSOM DOI 10.1287/msom.2024.0761), PDF pp. 46–47 (Online Supplement pp. 14–15), Lemma D.1
Human review
  • Endorsed by Shuze Chen · Oct 14, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 14, 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