Article
Keywords:
biorthogonalization; linear equations; biconjugate gradient method
Summary:
Lanczos’ method for solving the system of linear algebraic equations $Ax=b$ consists in constructing a sequence of vectors $x_k$ in such a way that $r_k=b-Ax_k \in r_0+A{\mathcal K}_{k}(A,r_0)$ and $r_k \perp {\mathcal K}_{k}(A^T,\widetilde{r}_0)$. This sequence of vectors can be computed by the BiCG (BiOMin) algorithm. In this paper is shown how to obtain the recurrences of BiCG (BiOMin) directly from this conditions.
References:
[Brezn–94] C. Brezinski, M. Redivo-Zaglia:
Treatment of near-breakdown in the CGS algorithm. Numerical Algorithms 7 (1994), 33–73.
DOI 10.1007/BF02141260 |
MR 1283334
[Fletch–76] R. Fletcher:
Conjugate gradient methods for indefinite systems. Numerical Analysis, Dundee, 1975, G. A. Watson (ed.), Vol. 506 of Lecture Notes in Mathematics, Springer, Berlin, 1976.
MR 0461857 |
Zbl 0326.65033
[Gutkn–97] M. H. Gutknecht:
Lanczos-type Solvers for Nonsymmetric Linear Systems of Equations. Technical Report TR-97-04, Swiss Center for Scientific Computing ETH-Zentrum, Switzerland, 1997.
MR 1489258 |
Zbl 0888.65030
[Lancz–50] C. Lanczos:
An iteration method for the solution of eigenvalue problem of linear differential and integral operators. J. Res. Nat. Bureau Standards 45 (1950).
DOI 10.6028/jres.045.026 |
MR 0042791
[Leary–80] D. P. O‘Leary:
The block conjugate gradient algorithm. Linear Algebra Appl. 99 (1980), 293–322.
MR 0562766
[Tichý–97] P. Tichý: Behaviour of BiCG and CGS algorithms. Mgr. thesis, Department of Numerical Mathematics, Faculty of Mathematics and Physics Praha, 1997.