An iterative method for the solution of nonlinear systems using the Faber polynomials for annular sectors

Conference · Sat Dec 31 04:00:00 EST 1994

OSTI ID:223850

Myers, N J ^[1]

---

1. Univ. of Durham (United Kingdom)

The author gives a hybrid method for the iterative solution of linear systems of equations Ax = b, where the matrix (A) is nonsingular, sparse and nonsymmetric. As in a method developed by Starke and Varga the method begins with a number of steps of the Arnoldi method to produce some information on the location of the spectrum of A. This method then switches to an iterative method based on the Faber polynomials for an annular sector placed around these eigenvalue estimates. The Faber polynomials for an annular sector are used because, firstly an annular sector can easily be placed around any eigenvalue estimates bounded away from zero, and secondly the Faber polynomials are known analytically for an annular sector. Finally the author gives three numerical examples, two of which allow comparison with Starke and Varga`s results. The third is an example of a matrix for which many iterative methods would fall, but this method converges.

• Research Organization: Front Range Scientific Computations, Inc., Boulder, CO (United States); USDOE, Washington, DC (United States); National Science Foundation, Washington, DC (United States)
• OSTI ID: 223850
• Report Number(s): CONF-9404305--Vol.1; ON: DE96005735
• Country of Publication: United States
• Language: English

Related Subjects

99 GENERAL AND MISCELLANEOUS
ASYMMETRY
ITERATIVE METHODS
MATRICES
NONLINEAR PROBLEMS
POLYNOMIALS