Multivariate Descartes' rule

From MaRDI portal





This paper studies polynomial systems \(P_1,\dots, P_k\) of real polynomials in \(k\) variables with fixed set of Newton polytopes \(\Delta_1, \Delta_2,\dots, \Delta_k\) (\(\Delta_i\) is the convex hull of the exponent vectors of monomials in \(P_i\)) and a fixed set of signs in the coefficients. The main questions are: What is the number of real zeros of system in a given open orthant? What is the number of real zeros with nonzero coordinates? How large can these number be? The authors guarantee a combinatorial lower bound for these quantities. The construction is based on subdivisions of Minkowski sums of the Newton polytopes \(\Delta_1, \Delta_2,\dots, \Delta_k\), the so called Viro's method (more precisely a version proposed by Sturmfels). The authors conjectured the bounds are actually upper bounds, this motivated a lot of interst and active research. A recent counterexample was announced by \textit{T. Y. Li} and \textit{X. Wang}. It is given by a 2-variable system \[ -1-x_1+x_2, \qquad -2-9x_1^3+ x^3_2+ 0.01 x^3_1 x^3_2. \]











This page was built for publication: Multivariate Descartes' rule

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2365253)