Methods of semiconjugate directions
Валерий Павлович Ильин · Russian Journal of Numerical Analysis and Mathematical Modelling · 2008
Methods of iterative solution in Krylov subspaces are considered for systems of linear algebraic equations with positive definite nonsymmetric real matrices A , where directing (correcting) vectors and residual vectors possess different properties of A -semiconjugacy or orthogonality. Variational and projective properties of some variants of such algorithms and the possibility to use dynamic preconditioners are described. Results of numerical experiments are presented.