Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 6.1, p. 20 — k-HL is universally ex-post IC-IR and a (k+3)-approximation to the optimal welfare for SOS valuations with k signal values

Proved
InterdepSOS.KHL.theorem_6_1

by mikedeng1 · Oct 15, 2026 · Mathlib 0df444a (Lean v4.33.1)

ex-post-incentive-compatibilityk-hl-mechanismp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1welfare-approximation

Consider a combinatorial auction with nnn agents, mmm items and a positive number kkk of signal values, with single-dimensional signals si∈{0,1,…,k−1}s_i\in\{0,1,\dots,k-1\}si​∈{0,1,…,k−1}, where the valuations viTv_{iT}viT​ satisfy the standing assumptions (nonnegative, vi∅≡0v_{i\emptyset}\equiv0vi∅​≡0, weakly increasing in every signal and strictly increasing in the own signal on nonempty bundles) and are SOS. Run the mechanism k-HL: with probability pRT=k−1k+3p_{RT}=\frac{k-1}{k+3}pRT​=k+3k−1​ run Random Threshold with ℓ\ellℓ uniform in {1,…,k−1}\{1,\dots,k-1\}{1,…,k−1}, otherwise Random Sampling with a uniformly random partition (A,B)(A,B)(A,B); the allocations are produced by arbitrary maximizing selectors that depend on the reports only through the proxy data. Then:

  1. for every threshold ℓ∈{1,…,k−1}\ell\in\{1,\dots,k-1\}ℓ∈{1,…,k−1}, Random Threshold with its payments viTˉi(s−i,si=ℓ−1)v_{i\bar T_i}(\mathbf s_{-i},s_i=\ell-1)viTˉi​​(s−i​,si​=ℓ−1) is ex-post IC and ex-post IR;
  2. for every B⊆[n]B\subseteq[n]B⊆[n], Random Sampling without payments is ex-post IC and ex-post IR;
  3. for every grid profile s\mathbf ss and every feasible allocation T∗T^*T∗,
∑iviTi∗(s)  ≤  (k+3)(pRT⋅1k−1∑ℓ=1k−1∑iviTˉiℓ(s)+(1−pRT)⋅12n∑B⊆[n]∑iviT~iB(s)).\sum_i v_{iT^*_i}(\mathbf s)\;\le\;(k+3)\Bigl(p_{RT}\cdot\frac1{k-1}\sum_{\ell=1}^{k-1}\sum_i v_{i\bar T^{\ell}_i}(\mathbf s)+(1-p_{RT})\cdot\frac1{2^n}\sum_{B\subseteq[n]}\sum_i v_{i\tilde T^{B}_i}(\mathbf s)\Bigr).i∑​viTi∗​​(s)≤(k+3)(pRT​⋅k−11​ℓ=1∑k−1​i∑​viTˉiℓ​​(s)+(1−pRT​)⋅2n1​B⊆[n]∑​i∑​viT~iB​​(s)).

Items 1–2 say that k-HL is universally ex-post IC-IR; item 3 says that its expected welfare at the truthful profile, measured with true values, is at least a 1/(k+3)1/(k+3)1/(k+3) fraction of the optimal welfare.

Formalization Note The approximation is stated against every feasible allocation instead of an optimum, which is equivalent and avoids a supremum. The paper's "signal space of size kkk" is made explicit as 1≤k1\le k1≤k; at k=0k=0k=0 the displayed mixing weight would be negative. For k=1k=1k=1, pRT=0p_{RT}=0pRT​=0 and k-HL is Random Sampling.

Preamble
import Mathlib
import Definitions.Def_InterdepSOS_KHL_Setting
import Definitions.Def_InterdepSOS_KHL_Mechanisms
Formal statement
namespace InterdepSOS.KHL
theorem theorem_6_1 {n m : ℕ} (k : ℕ) (hk : 1 ≤ k) (v : Valuation n m)
    (hv : IsCombValuation k v) (hsos : ∀ i T, IsSOS k (v i T))
    (rtSel : ℕ → Signal n → Allocation n m) (hrt : IsRTSelector k v rtSel)
    (rsSel : Finset (Fin n) → Signal n → Allocation n m) (hrs : IsRSSelector k v rsSel) :
    (∀ ℓ ∈ Finset.Icc 1 (k - 1),
        IsExPostIC k v (rtAlloc rtSel ℓ) (rtPay v rtSel ℓ) ∧
        IsExPostIR k v (rtAlloc rtSel ℓ) (rtPay v rtSel ℓ)) ∧
    (∀ B : Finset (Fin n),
        IsExPostIC k v (rsAlloc rsSel B) 0 ∧ IsExPostIR k v (rsAlloc rsSel B) 0) ∧
    ∀ s : Signal n, InGrid k s → ∀ Tstar : Allocation n m, IsAllocOn Finset.univ Tstar →
      welfare v Tstar s ≤ ((k : ℝ) + 3) * khlWelfare k v rtSel rsSel s := by sorry
end InterdepSOS.KHL
Source
Eden, Feldman, Fiat, Goldner, Karlin, Combinatorial Auctions with Interdependent Valuations: SOS to the Rescue, arXiv:1903.08384v2 (2019), p. 20, Theorem 6.1
Human review
  • Endorsed by Shuze Chen · Oct 15, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 15, 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