BI-CGSTAB: A fast and smoothly converging variant of BI-CG for the solution of nonsymmetric linear systems. (English) Zbl 0761.65023
In order to improve the convergence properties of the conjugate gradients-squared method, the author suggests making use of preconditioned matrices. Four examples of symmetric and nonsymmetric matrices are considered where modified incomplete Cholesky decomposition or incomplete LU-factorization is used. The calculation results show that the residual in this case decreases considerably faster and that this decrease is almost monotonous.
Reviewer: Y.Shokin (Krasnoyarsk)
MSC:
65F10 | Iterative numerical methods for linear systems |
65F35 | Numerical computation of matrix norms, conditioning, scaling |