Preconditioned Iterative Methods for Homotopy Curve Tracking

From MaRDI portal





Homotopy algorithms represent a class of methods for solving systems of nonlinear equations that are globally convergent with probability one. These algorithms are based on the construction of an appropriate homotopy map and then the tracking of a curve in the zero set of this homotopy map. The fundamental computation steps in this algorithm can be transformed into a linear algebra problem based on the computation of the kernel of the homotopy Jacobian matrix. Problems with a large but sparse Jacobian matrix are considered in the paper. The curve-tracking algorithms used require the solution of a series of very special systems, in particular \((n+1)\times (n+1)\) in general nonsymmetric systems, which have a leading symmetric indefinite \(n\times n\) submatrix. Furthermore, the last row of each such system may be chosen almost arbitrarily. The authors seek to take advantage of these special properties. The iterative methods studied in the paper include \textit{E. J. Craig's} variant of conjugate gradient algorithm and the SYMMLQ algorithm for symmetric indefinite problems. The effectiveness of various preconditioning strategies in this context are also investigated and several choices for the last row of the systems to be solved are explored in the paper.





Describes a project that uses

Uses Software






This page was built for publication: Preconditioned Iterative Methods for Homotopy Curve Tracking

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3990902)