Discrete geometry for algebraic elimination

From MaRDI portal





This article surveys efficient methods based on the toric (or sparse) resultant for computing all isolated solutions of an arbitrary system of \(n\) polynomials in \(n\) unknowns. The author constructs matrix formulae which yield nontrivial multiples of the resultant thus reducing root-finding to the eigendecomposition of a square matrix or to multivariate polynomial factorization. Other elimination problems treated by resultants include the determination of predicates and the implicitization of parametric varieties. The author describes briefly the main steps in the theory of toric elimination which generalizes the results of the classical theory to polynomials described by their supports, thus exploiting their spareness. Then the construction of toric resultant matrices of Sylvester-type is presented. The solution of arbitrary algebraic system is reduced to numerical linear algebra, thus yielding methods which avoid any issues of convergence. Finally, the author gives examples with the incremental algorithm and attempts a comparison to Gröbner basis software on the systems of the cyclic \(n\)-family. All algorithms discussed are implemented either in Maple and/or in C and are available through the author's webpage. For the entire collection see [Zbl 1008.00013].





Describes a project that uses

Uses Software






This page was built for publication: Discrete geometry for algebraic elimination

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4418122)