Integer H-representations in
DefinitionHirschDim4_Settingdiscrete-geometrylinear-optimizationpolytopes
This file fixes the exact-data convention for the dimension- Hirsch mission. An integer matrix and an integer vector are read into real H-representation data: intRows A is the list of row normals and intRhs b the right-hand sides , so that Hirsch.Hpoly (intRows A) (intRhs b) is the polytope . Integer data are equivalent to a rational polytope (clearing denominators), matching the benchmark's request for an explicit rational counterexample with an exact integer H-representation. These coercion helpers are reused by the goal theorem.
Definition code
import Mathlib
open scoped RealInnerProductSpace
namespace HirschDim4
noncomputable section
/-- The rows of an integer matrix `A`, viewed as normal vectors in `ℝ⁴`. Together with
`intRhs` this turns an exact integer H-representation into the platform's real `Hpoly` data. -/
def intRows {n : ℕ} (A : Matrix (Fin n) (Fin 4) ℤ) :
Fin n → EuclideanSpace ℝ (Fin 4) :=
fun i => WithLp.toLp 2 (fun j => (A i j : ℝ))
/-- An integer right-hand side, viewed in `ℝ`. -/
def intRhs {n : ℕ} (b : Fin n → ℤ) : Fin n → ℝ := fun i => (b i : ℝ)
end
end HirschDim4
Source
Epoch AI, FrontierMath: Open Problems, The Hirsch Conjecture in Dimension 4, https://epoch.ai/frontiermath/open-problems/hirsch-conjecture (submission format: integer matrix A and integer vector b).
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.