Skip to main content
Log in

Partitioned matrices and Seidel convergence

  • Published:
Numerische Mathematik Aims and scope Submit manuscript

Summary

Without using spectral resolution, an elementary proof of convergence of Seidel iteration. The proof is based on the lemma (generalizing a lemma of P. Stein): If (A+A *)−B *(A+A *)B>0, whereB=−(P+L) −1 R,A=P+L (Lower)+R (upper), then Seidel iteration ofAX=Y 0 converges if and only ifA+A *>0. This lemma has as corollaries not only the well-known results of E. Reich and Stein, but also applications to a matrix that can be far from symmetric, e.g.M=[A ij ] 21 , whereA 21=−A *12 ,A 11,A 22 are invertible;A 11 +A *11 =A22+A *22 ; and the proper values ofA −112 A 11,A *−112 A 22 are in the interior of the unit disk.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. Reich, Edgar: On the convergence of the classical iterative method of solving linear simultaneous equations. Ann. Math. Stat.20, 448–451 (1949).

    Google Scholar 

  2. Stein, P.: The convergence of Seidel iterants of nearly symmetric matrices. Math. Tables and Aids to Computation5, 237–240 (1951).

    Google Scholar 

  3. Varga, R. S.: Matrix iterative analysis. New York: Prentice Hall 1963.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Additional information

Supported under NSF GP 32527.

Supported under NSF GP 8758.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Brenner, J.L., de Pillis, J. Partitioned matrices and Seidel convergence. Numer. Math. 19, 76–80 (1972). https://doi.org/10.1007/BF01395932

Download citation

  • Received:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF01395932

Keywords

Navigation