Title:Minimal determinantal representations of bivariate polynomials

Abstract:For a generic bivariate polynomial $p$ of degree $n$ we introduce a simple and fast numerical algorithm for the construction of $n\times n$ matrices $A$, $B$, and $C$ such that $\det(A+xB+yC) = p(x,y)$. This is the smallest sufficient size to represent a bivariate polynomial of degree $n$ and we give the first algorithm that can compute matrices of this size for almost all bivariate polynomials. The existence of such symmetric matrices for any bivariate polynomial of degree n was established by Dixon in 1902, but, up to now, no simple numerical construction has been found, even if we allow that the matrices are not symmetric. The obtained representations may be used to efficiently numerically compute the roots of a system of two bivariate polynomials of small degree via the eigenvalues of a two-parameter eigenvalue problem. We show that the new representation speeds up the computation considerably.