Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 12, p. 28 — the powers-of-δ optimum of (4), scaled to (δ^(2γ̄+1) x̂, δ^(γ̄+1) ŷ), is feasible for (3)

Proved
NestedLogitVariants.PowersDelta.theorem_12

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

approximation-algorithmsassortment-optimizationnested-logitp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Throughout, the instance satisfies the standing assumptions of §1, there is at least one product (n≥1n \ge 1n≥1), γˉ=max⁡i∈Mγi\bar\gamma = \max_{i \in M} \gamma_iγˉ​=maxi∈M​γi​ satisfies γˉ>1\bar\gamma > 1γˉ​>1 (the standing assumption of §6, p. 25), and δ>1\delta > 1δ>1. Let (x^,y^)(\hat x, \hat y)(x^,y^​) be an optimal solution of problem (4) when the candidate collection of nest iii is {S^il:l=liL,…,liU}∪{∅}\{\hat S_{il} : l = l^L_i, \dots, l^U_i\} \cup \{\emptyset\}{S^il​:l=liL​,…,liU​}∪{∅}, and the assortments S^il\hat S_{il}S^il​ have the two properties of p. 28: whenever problem (15) at level lll is feasible, S^il\hat S_{il}S^il​ is feasible for it and δ∑j∈S^ilrijvij≥G^il\delta \sum_{j \in \hat S_{il}} r_{ij} v_{ij} \ge \hat G_{il}δ∑j∈S^il​​rij​vij​≥G^il​. Then

(δ2γˉ+1x^, δγˉ+1y^) is a feasible solution to problem (3),\big(\delta^{2\bar\gamma + 1} \hat x,\ \delta^{\bar\gamma + 1} \hat y\big) \text{ is a feasible solution to problem (3)},(δ2γˉ​+1x^, δγˉ​+1y^​) is a feasible solution to problem (3),

that is, v0 δ2γˉ+1x^≥∑i∈Mδγˉ+1y^iv_0\, \delta^{2\bar\gamma+1} \hat x \ge \sum_{i \in M} \delta^{\bar\gamma+1} \hat y_iv0​δ2γˉ​+1x^≥∑i∈M​δγˉ​+1y^​i​, and δγˉ+1y^i≥Vi(Si)γi(Ri(Si)−δ2γˉ+1x^)\delta^{\bar\gamma+1} \hat y_i \ge V_i(S_i)^{\gamma_i} \big(R_i(S_i) - \delta^{2\bar\gamma+1} \hat x\big)δγˉ​+1y^​i​≥Vi​(Si​)γi​(Ri​(Si​)−δ2γˉ​+1x^) for every nest iii and every assortment Si⊆NS_i \subseteq NSi​⊆N.

Combined with Theorem 1, this shows that restricting nest iii to the at most 2+log⁡δ(viU/viL)2 + \log_\delta(v^U_i / v^L_i)2+logδ​(viU​/viL​) candidates {S^il}∪{∅}\{\hat S_{il}\} \cup \{\emptyset\}{S^il​}∪{∅} loses at most a factor δ2γˉ+1\delta^{2\bar\gamma+1}δ2γˉ​+1 of the optimal expected revenue, for the general nested logit model with nest dissimilarity parameters above one and arbitrary nonnegative no-purchase weights in the nests.

Formalization Note The assortments S^il\hat S_{il}S^il​ enter as an arbitrary family with the two properties of p. 28 (IsPowersFamily): the statement holds for every such family, which includes the one Proposition 15 constructs. A level with no feasible assortment imposes nothing on S^il\hat S_{il}S^il​. Optimality of (x^,y^)(\hat x, \hat y)(x^,y^​) for (4) is feasibility plus minimality of x^\hat xx^ among feasible points. γˉ\bar\gammaγˉ​ is passed as the greatest element of {γi}\{\gamma_i\}{γi​}, so there is at least one nest. The pins of the shared model (vij>0v_{ij} > 0vij​>0, rij≥0r_{ij} \ge 0rij​≥0, γi>0\gamma_i > 0γi​>0) are disclosed in the definition item.


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_NestedLogitVariants_PowersDelta_Levels
Formal statement
namespace NestedLogitVariants.PowersDelta

/-- Theorem 12, p. 28. Let `(x̂, ŷ)` be an optimal solution of problem (4) with the candidate
collections `{Ŝ_il : l = l^L_i, …, l^U_i} ∪ {∅}`. Then `(δ^{2γ̄+1} x̂, δ^{γ̄+1} ŷ)` is feasible for
problem (3), where `γ̄ = max_i γ_i > 1` and `δ > 1`. Stated for every family `Ŝ_il` with the two
properties of p. 28 (`IsPowersFamily`). -/
theorem theorem_12 {ι : Type*} [Fintype ι] [DecidableEq ι] {n : ℕ} [NeZero n] (I : Instance ι n)
    (hI : I.Standing) (γbar : ℝ) (hγbar : IsGreatest (Set.range I.γ) γbar) (hγbar1 : 1 < γbar)
    (δ : ℝ) (hδ : 1 < δ)
    (Sh : ι → ℤ → Finset (Fin n)) (hSh : IsPowersFamily I δ Sh)
    (xh : ℝ) (yh : ι → ℝ) (hopt : LP4Optimal I (powersCandidates I δ Sh) xh yh) :
    LP3Feasible I (δ ^ (2 * γbar + 1) * xh) (δ ^ (γbar + 1) • yh) := by sorry

end NestedLogitVariants.PowersDelta
Source
Davis, Gallego, Topaloglu, Assortment optimization under variants of the nested logit model, revised manuscript of June 18, 2013 (published Oper. Res. 62(2), 2014, DOI 10.1287/opre.2014.1256), p. 28, Theorem 12; proof in Appendix A.6, pp. 53–54
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