Summary
The asymptotic convergence of the coordinate overrelaxation method for computing the smallest eigenvalue and corresponding eigenvector of the general eigenvalue problem (A−λB)x=0 is investigated. A certain analogy between the asymptotic behaviour of this algorithm and the successive overrelaxation method is developed in deriving the appropriate iteration matrix. By the aid of this matrix it is shown that asymptotic convergence of the process towards the eigenvector of the smallest eigenvalue is guaranteed for all relaxation factors ω ε (0,2).The optimal choice of ω can be discussed theoretically in case thatA andB have “propertyA”. The theoretical facts are illustrated by two examples originating from the problem of the vibrating string using spline functions.
Similar content being viewed by others
References
Bender, C. F., Shavitt, I.: An iterative procedure for the calculation of the lowest real eigen value and eigenvector of a nonsymmetric matrix. J. Computational Physics6, 146–149 (1970)
Bradbury, W. W.: Fletcher, R.: New iterative methods for solution of the eigenproblem. Numer. Math.9, 259–267 (1966)
Faddejew, D. K., Faddejewa, W. N.: Numerische Methoden der linearen Algebra. München-Wien: Oldenbourg 1964. Computational methods of linear algebra. San Franzisco: Freeman & Co. 1963
Falk, S.: Ein einfaches Iterationsverfahren zur Bestimmung der Eigenwerte eines hermiteschen (reellsymmetrischen) Matrizenpaares. Acta Technica Academia Scientiarum Hungaricae73, 327–334 (1972)
Falk, S.: Berechnung von Eigenwetten und Eigenvektoren normale Matrizenpaare durch Ritz-Iteration. ZaMM53, 73–91 (1973)
Holm, S.: Coordinate overrelaxation methods for the eigenproblem. Report UMINF-33.73, Umea University, 1973
Kahan, W.: Gauss-Seidel methods for solving large systems of linear equations. Doctoral thesis, University of Toronto, Toronto, Canada 1958
Kahan, W.: Relaxation methods for an eigenproblem. Technical Report CS-44, Computer Science Department Stanford University, 1966
Keller, H. B.: On the solution of singular and semidefinite linear systems by iteration. J. SIAM Numer. Anal. Ser. B,2, 281–290 (1965)
Nesbet, R. K.: Algorithm for diagonalization of large matrices. J. Chem. Physics43, 311–312 (1965)
Nisbet, R. M.: Acceleration of the convergence in Nesbet's algorithm for eigenvalues and eigenvectors of largee matrices. J. Computational Physics10 614–619 (1972)
Ruhe, A.: SOR-methods for the eigenvalue problem. Report UMINF-37.73, Umea University, 1973
Schwarz, H. R., Rutishauser, H., Stiefel, E.: Numerik symmettrischer Matrizen, 2. Aufl. Stuttgart: Teubner 1972. Numerical analysis of symmetric matrices. Englewood Cliffs: Prentice-Hall 1973
Schwarz, H. R.: The eigenvalue problem(A−νB)x=0 for symmetric matrices of high order. Computer methods in applied mechanics and engineering3, 11–28 (1974)
Shavitt, I.: Modification of Nesbet's algorithm for the iterative evaluation of eigenvalues and eigenvectors of large matrices. J. Computational Physics6, 124–130 (1970)
Shavitt, I., Bender, C. F., Pipano, A., Hosteny, R. P.: The iterative calculation of several of the lowest or highest eigenvalues and corresponding eigenvectors of very large symmetric matrices. J. Computational Physics11, 90–108 (1973)
Stewart, G. W.: On the sensitivity of the eigenvalue problemAx=νBx SIAM J. Numer. Analysis9, 669–686 (1972)
Wilkinson, J. H.: The algebraic eigenvalue problem. Oxford: Clarendon Press 1965
Wilkinson, J. H., Reinsch, C.: Handbook for automatic computation, vol. II, Linear Algebra. Berlin-Heidelberg-New York: Springer 1971
Young, D. M.: Iterative solution of large linear systems. New York: Academic Press 1971
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Schwarz, H.R. The method of coordinate overrelaxation for (A−λB)x=0. Numer. Math. 23, 135–151 (1974). https://doi.org/10.1007/BF01459947
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF01459947