AFTD Auto-Formalizing Theoretical Domains

is_rank_score_veto_core

def

Given a nonempty finite voter type V, a finite candidate type C with decidable equality, a scoring vector s : Fin (Fintype.card C) → ℚ, an election profile prof : V → LinearOrder C, and a candidate a in C, is_rank_score_veto_core s prof a asserts that has_fractional_matching prof (fun _ => 1 / (Fintype.card V : ℝ)) (fun c => ((rank_score_candidate_weight s prof c : ℚ) : ℝ)) a.

Statement

def is_rank_score_veto_core {V C : Type*} [Fintype V] [Nonempty V] [Fintype C] [DecidableEq C]
    (s : Fin (Fintype.card C) → ℚ) (prof : V → LinearOrder C) (a : C) : Prop
Defined in
arXiv:2610.26535 Definition 2, Section 2.3
Verified
24 Oct 2026
Axioms
none listed
Built on 2
  • has_fractional_matchingGiven finite types V and C, a profile prof : V → LinearOrder C assigning a linear order to each voter v in V, voter weights p : V → ℝ,…
  • rank_score_candidate_weightGiven a nonempty finite type of voters V, a finite type of candidates C with decidable equality, a scoring function s : Fin (Fintype.card…
Used by 2

Read back from the Lean

Given finite types V (with V nonempty) and C (with decidable equality), a scoring vector s : Fin (Fintype.card C) → ℚ, a preference profile prof : V → LinearOrder C, and an alternative a : C, is_rank_score_veto_core s prof a asserts the existence of a fractional matching (weight matrix) w : V → C → ℝ such that: 1. Every weight is non-negative: for all v : V and c : C, 0 ≤ w v c. 2. Positive weights only go to alternatives weakly worse than (or equal to) a: for all v : V and c : C, if 0 < w v c, then c ≤ a according to the linear order prof v ((prof v).le c a). 3. The row sums (marginal over candidates for each voter v) are all equal to the uniform share 1 / (Fintype.card V : ℝ). 4. The column sums (marginal over voters for each candidate c) equal the normalized rank-score weight of candidate c under scoring vector s and profile prof, cast to ℝ: ∑ v : V, w v c = ((∑ v : V, s (candidate_rank (prof v) c)) / ((Fintype.card V : ℚ) * ∑ i : Fin (Fintype.card C), s i) : ℝ). Here, candidate_rank ord c is the number of alternatives c' : C strictly greater than c in ord (i.e., alternatives preferred to c).

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

Lean source view module on GitHub

/-- Membership of candidate a in the (p^unif, q^s)-veto core. -/
def is_rank_score_veto_core {V C : Type*} [Fintype V] [Nonempty V] [Fintype C] [DecidableEq C]
    (s : Fin (Fintype.card C) → ℚ) (prof : V → LinearOrder C) (a : C) : Prop :=
  has_fractional_matching prof
    (fun _ => 1 / (Fintype.card V : ℝ))
    (fun c => ((rank_score_candidate_weight s prof c : ℚ) : ℝ))
    a
See it in the graphDiscuss this resultChallenge it