The granularity of parallel homotopy algorithms for polynomial systems of equations

From MaRDI portal





In recent years several authors have considered the application of homotopies and numerical continuation methods to the problem of finding all of the finite zeroes in \(C^ n\) of a complex system of polynomial equations. For examples of references we suggest the following papers and the references therein: \textit{S.-N. Chow}, \textit{J. Mallet-Paret} and \textit{J. A. Yorke} [Math. Comp. 32, 887-899 (1978; Zbl 0398.65029)]; \textit{C. B. Garcia} and \textit{W. I. Zangwill} [Math. Oper. Res. 4, 1-14 (1979; Zbl 0408.90086)]; \textit{T. Y. Li}, \textit{T. Sauer} and \textit{J. A. Yorke} [Numer. Math. 51, 481-500 (1987; Zbl 0656.65055)]. One of the drawbacks of the homotopy method in this context is the large number of curves which need to be traversed. Generically, this is the product of the degrees of the component maps. However, the situation lends itself very well to parallelizing. In the present paper the efficacy of parallel computing for this problem is explored. Very interesting data is given concerning a large number of practical problems which were computed on a variety of machines.





Describes a project that uses

Uses Software






This page was built for publication: The granularity of parallel homotopy algorithms for polynomial systems of equations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4733310)