polygon_v2
For a natural number r, polygon_v2 r is V₂(r), defined as the supremum of the set of vertex counts of convex polygons in the plane admitting a real positive semidefinite slack factorization of size at most r.
Statement
noncomputable def polygon_v2 (r : ℕ) : ℕ
- Defined in
- arXiv:2610.04498 Sec. 1
- Verified
- 23 Oct 2026
- Axioms
- none listed
- Built on 3
- is_convex_polygon_slack_matrixA real matrix S : Matrix (Fin n) (Fin n) ℝ is the slack matrix of a convex n-gon in the plane (is_convex_polygon_slack_matrix n S) if 3 ≤…
- matrix_has_psd_factorizationFor types m and n, a real matrix M : Matrix m n ℝ has a positive semidefinite (PSD) factorization of size r (for a natural number r) if…
- polygon_psd_vertex_countFor a natural number r, polygon_psd_vertex_count r is the set of natural numbers n such that there exists an n × n real matrix S that is…
- Used by 2
- polygon_v2_four_eq_eight_iff_no_largerAssuming an 8-gon with size-4 positive semidefinite slack factorization exists and the set of realizable vertex counts is bounded above,…
- polygon_v2_four_ge_eightIf there exists a convex octagon in the plane admitting a real positive semidefinite slack factorization of size 4, and the set of…
Read back from the Lean
polygon_v2 is a function mapping a natural number r to the supremum (in ℕ) of the set of natural numbers n such that there exists an n × n real matrix S satisfying:
1. S is the slack matrix of an n-gon in ℝ², meaning 3 ≤ n and there exist an injective vertex map v : Fin n → Fin 2 → ℝ and affine functionals ℓ : Fin n → (Fin 2 → ℝ) →ᵃ[ℝ] ℝ such that:
- The extreme points of the convex hull of the vertices {v i | i} are exactly the vertices themselves.
- The affine span of the vertices is the full space ℝ² (the polygon is 2-dimensional).
- The convex hull of the vertices equals the polyhedron {x | ∀ i, 0 ≤ ℓ i x} defined by the n affine inequalities.
- For every facet i, there are at least two distinct vertices j₁ ≠ j₂ on the boundary ℓ i (v j₁) = 0 and ℓ i (v j₂) = 0.
- The matrix entries are the slacks: S i j = ℓ i (v j) for all i, j.
2. S has a positive semidefinite (PSD) factorization of size r, meaning there exist families of r × r real matrices A, B : Fin n → Matrix (Fin r) (Fin r) ℝ such that each A i is positive semidefinite, each B j is positive semidefinite, and S i j = Matrix.trace (A i * B j) for all i, j.
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
/-- The maximum number of vertices V₂(r) of a convex polygon in the plane admitting a real positive semidefinite slack factorization of size at most r. -/ noncomputable def polygon_v2 (r : ℕ) : ℕ := sSup (polygon_psd_vertex_count r)