Resultant and squarefree coefficient of a connected balanced binomial system
ProvedResultantVNP.connected_binomial_resultantgraph-theorypolynomialsresultants
Let f and g map {1,…,n} to itself, with n ≥ 1. Suppose the directed multigraph with two arcs i → f(i), g(i) at each vertex has indegree two at each vertex and is connected after forgetting arc directions. For any normalized multivariate resultant system, let P(t) be the resultant of the quadratic forms x_i² − t_i x_{f(i)}x_{g(i)}. There is an integer c ≥ 1 such that
In particular, this squarefree coefficient is a negative integer. Loops and repeated arcs are allowed; all arcs are counted with multiplicity.
Preamble
import Mathlib import Definitions.Def_ResultantVNP_ResultantSystem
Formal statement
theorem ResultantVNP.connected_binomial_resultant (𝓡 : ResultantVNP.ResultantSystem) {n : ℕ} (hn : 1 ≤ n)
(f g : Fin n → Fin n)
(hb : ∀ v,(Finset.univ.filter (fun u => f u = v)).card +
(Finset.univ.filter (fun u => g u = v)).card = 2)
(hc : ∀ u v,Relation.ReflTransGen
(fun a b => f a = b ∨ g a = b ∨ f b = a ∨ g b = a) u v) :
let P : MvPolynomial (Fin n) ℂ :=
MvPolynomial.aeval (fun v : ResultantVNP.CoeffVar (fun _ : Fin n => 2) =>
(MvPolynomial.X v.1 ^ 2 - MvPolynomial.C (MvPolynomial.X v.1) *
MvPolynomial.X (f v.1) * MvPolynomial.X (g v.1) :
MvPolynomial (Fin n) (MvPolynomial (Fin n) ℂ)).coeff v.2.1)
(MvPolynomial.map (Int.castRingHom ℂ) (𝓡.res (fun _ => 2)))
∃ c : ℕ, 1 ≤ c ∧ P = (1 - ∏ i : Fin n,(MvPolynomial.X i : MvPolynomial (Fin n) ℂ)) ^ c ∧
P.coeff (Finsupp.indicator Finset.univ (fun _ _ => 1)) = -(c : ℂ) := by sorrySource
The binomial-system step in the proof of Theorem 3.5 in Bhargav–Curticapean–Jindal–Shukla, Multivariate Resultants: Complexity and Applications to Real Feasibility, STOC 2027 paper #1103; generalized to any connected balanced two-outdegree directed multigraph.