DOI QR코드

DOI QR Code

AN ITERATIVE METHOD FOR SYMMETRIC INDEFINITE LINEAR SYSTEMS

  • Walker, Homer-F. (Department of Mathematical Sciences Worcester Polytechnic Institute) ;
  • Yi, Su-Cheol (Department of Applied Mathematics Changwon National University)
  • Published : 2004.04.01

Abstract

For solving symmetric systems of linear equations, it is shown that a new Krylov subspace method can be obtained. The new approach is one of the projection methods, and we call it the projection method for convenience in this paper. The projection method maintains the residual vector like simpler GMRES, symmetric QMR, SYMMLQ, and MINRES. By studying the quasiminimal residual method, we show that an extended projection method and the scaled symmetric QMR method are equivalent.

Keywords

GMRES;MINRES;SYMMLQ;symmetric QMR;and Krylov subspace methods.

References

  1. ORNL/TM-12754 A new Krylov subspace method for symmetric indefinite linear systems
  2. Murray Hill Bell Labs Simplifications of the nonsymmetric Lanczos process and a new algorithm for Hermitian indefinite linear systems R.W.Freund;H.Zha
  3. Matrix Computations G.H.Golub;C.F.Van Loan
  4. SIAM J. Numer. Anal. v.12 Solution of sparse indefinite systems of linear equations C.C.Paige;M.A.Saunders https://doi.org/10.1137/0712047
  5. Numer. Lin. Alg. Appl. v.1 A simple GMRES H.F.Walker;Lu Zhou https://doi.org/10.1002/nla.1680010605
  6. Tech. Rep. 91-26 A quasi-minimal residual squared algorithm for non Hermitian linear systems R.W.Freund;H.Szeto
  7. SIAM J. Sci. Stat. Comput. v.7 GMRES:A generalized minimal residual algorithm for solving nonsymmetric linear systems Y.Saad;M.H.Schultz https://doi.org/10.1137/0907058
  8. Numer. Math. v.60 QMR:a quasi-minimal residual method for non-Hermitian linear systems R.W.Freund;N.M.Nachtigal https://doi.org/10.1007/BF01385726