Abstract
We show that every n-dimensional orthogonal matrix can be factored into O(n 2) Jacobi rotations (also called Givens rotations in the literature). It is well known that the Jacobi method,wh ich constructs the eigen-decomposition of a symmetric matrix through a sequence of Jacobi rotations,is slower than the eigenvalue algorithms currently used in practice,but is capable of computing eigenvalues, particularly tiny ones,t o a high relative accuracy. The above decomposition, which to the best of our knowledge is new,s hows that the infinite-precision nondeterministic Jacobi method (in which an oracle is invoked to obtain the correct rotation angle in each Jacobi rotation) can construct the eigen-decomposition with O(n 2) Jacobi rotations. The complexity of the nondeterministic Jacobi algorithm motivates the efforts to narrow the gap between the complexity of the known Jacobi algorithms and that of the nondeterministic version. Speeding up the Jacobi algorithm while retaining its excellent numerical properties would be of considerable interest. In that direction,we describe,as an example,a variant of the Jacobi method in which the rotation angle for each Jacobi rotation is computed (in closed-form) through 1-dimensional optimization. We also show that the computation of the closed-form optimal solution of the 1-dimensional problems guarantees convergence of the new method.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Arnoldi W.E. The principle of minimized iteration in the solution of the matrix eigenproblem. Quart. Appl. Math. 9 (1951) 17–29.
Brent R.P. and Luk F.T. The solution of singular value and symmetric eigenvalue problems on multiprocessor arrays. SIAM J. Sci. Statist. Comput. 6 (1985) 69–84.
Burnside W.S. and Panton A.W. The Theory of Equations,Dublin University Press Series, Dublin, Ireland, 1892.
Crouzeix M.,Philippe B. and Sadkane M. The Davidson method. SIAMJ. Sci. Comput. 15 (1994) 62–76.
Cullum J. and Willoughby R.A. Lanczos Algorithms for Large Symmetric Eigenvalue Computations. Vol. I, Birkhäuser, Boston, MA,1985.
Cullum J. and Willoughby R.A. Lanczos Algorithms for Large Symmetric Eigenvalue Computations. Vol. II, Birkhäuser, Boston, MA,1985.
Cullum J. and Willoughby R.A. A practical procedure for computing eigenvalues of large sparse nonsymmetric matrices. In: Cullum J.,Willoughby R.A. (Eds.),Large Scale Eigenvalue Problems, Amsterdam, North-Holland,1986, pp. 193–240.
Cuppen J.J.M. A divide and conquer method for the symmetric eigenproblem. Numer. Math. 36 (1981) 177–195.
Davidson E.R. The iterative calculation of a fewof the lowest eigenvalues and corresponding eigenvectors of large real symmetric matrices. J. Comput. Phys. 17 (1975) 87–94.
Demmel J.W. Applied Numerical Linear Algebra. SIAM, Philadelphia, PA, 1997.
Demmel J.W. and Veselić K. Jacobi’s method is more accurate than QR. SIAM J. Matrix Anal. Appl. 13 (1992) 1204–1246.
Dongarra J.J. and Sorensen D.C. A fully parallel algorithm for the symmetric eigenvalue problem. SIAM J. Sci. Statist. Comput. 8 (1987) 139–154.
Francis J.G.F. The QR transformation: a unitary analogue to the LR transformation. Parts I and II Comput. J. 4 1961 265–272,265–272.
Givens W. Numerical computation of the characteristic values of a real symmetric matrix. Oak Ridge National Lab.,ORNL-1574,1954.
Golub G.H. Some uses of the Lanczos algorithm in numerical linear algebra. In: Miller J.J.H. (Ed.), Topics in Numerical Analysis, A cademic Press, New York,1973, pp. 173–184.
Golub G.H. and Van Loan C.F. Matrix Computations. The Johns Hopkins University Press,Balt imore, 1996.
Golub G.H. and van der Vorst H.A. Eigenvalue computation in the 20th century. J. Comput. Appl. Math. 123 (2000) 35–65.
Greenstadt J. A method for finding roots of arbitrary matrices. Math. Tables Aids Comput. 9 (1955) 47–52.
Gregory R.T. Computing eigenvalues and eigenvectors of a symmetric matrix on the ILLIAC. Math. Tables Aids Comput. 7 (1953) 215–220.
Gu M. and Eisenstat S.C. A divide-and-conquer algorithm for the symmetric tridiagonal eigenproblem. SIAM J. Matrix Anal. Appl. 16 (1995) 172–191.
Henrici P. On the speed of convergence of cyclic and quasicyclic Jacobi methods for computing the eigenvalues of Hermitian matrices. SIAM J. Appl. Math. 6 (1958) 144–162.
Hessenberg K. Aufl¨osung linearer Eigenwertaufgaben met Hilfe der Hamilton- Cayleyschen Gleichung. Ph.D. Thesis,Technische Hochschule, Darmstadt, 1940.
Hoffmann W. and Parlett B.N. A newpr oof of global convergence for the tridiagonal QL algorithm. SIAM J. Numer. Anal. 15 (1978) 929–937.
Householder A.S. The Theory of Matrices in Numerical Analysis. Dover, New York,1964.
Jacobi C.G.J. Über ein leichtes Verfahren, die in der Theorie der Säcularstőrungen vorkommenden Gleichungen numerisch aufzulősen. J. Reine Angew. Math. 30 (1846) 51–94.
Krylov A.N. O cislennom resenii uravnenija, kotorym v techniceskih voprasah opredeljajutsja castoty malyh kolebanii material’nyh sistem. Izv. Akad. Nauk SSSR. Ser. Fiz.-Mat. 4 (1931) 491–539.
Kublanovskaja V.N. On some algorithms for the solution of the complete eigenvalue problem. Zh. Vych. Mat. 1 (1961) 555–570.
Lanczos C. An iteration method for the solution of the eigenvalue problem of linear differential and integral operators. J. Res. Natl. Bur. Stand 45 (1950) 225–280.
Li T.-Y. and Zeng Z. Homotopy-determinant algorithm for solving nonsymmetric eigenvalue problems. Math. Comput. 59 (1992) 483–502.
Liu S.H. and Golub G.H. Homotopy method for the numerical solution of the eigenvalue problem of self-adjoint partial differential operators. Numer. Algorithms 10 (1995) 363–378.
Morgan R.B. and Scott D.S. Generalizations of Davidson’s method for computing eigenvalues of sparse symmetric matrices. SIAM J. Sci. Statist. Comput. 7 (3) (1986) 817–825.
Olsen J.,J orgensen P. and Simons J. Passing the one-billion limit in full configuration-interaction (FCI) calculations. Chem. Phys. Lett. 169 (1990) 463–472.
[33] Ostrowski A.M. On the convergence of the Rayleigh quotient iteration for the computation of characteristic roots and vectors I-VI. Arch. Rational. Mesh. Anal. I 1:233–241,I I 2:233–241,I II 3:233–241,I V 3:233–241,V 3:233–241,V I 4:233–241,233–241.
Paige C.C. The computation of eigenvalues and eigenvectors of very large matrices. Ph.D. Thesis,University of London, 1971.
Parlett B.N. The Symmetric Eigenvalue Problem. Prentice-Hall, En glewood Cliffs,N J, 1980.
Pope D.A. and Tompkins C. Maximizing functions of rotations: experiments concerning speed of diagonalization of symmetric matrices using Jacobi’s method. J. ACM 4 (1957) 459–466.
Ruhe A. Implementation aspects of band Lanczos algorithms for computation of eigenvalues of large sparse symmetric matrices. Math. Comput. 33 (1979) 680–687.
Saad Y. Variations on Arnoldi’s method for computing eigenelements of large unsymmetric matrices. Linear Algebra Appl. 34 (1980) 269–295.
Schönhage A. Zur Konvergenz des Jacobi-Verfahrens. Numer. Math. 3 (1961) 374–380.
Sleijpen G.L.G. and van der Vorst H.A. A Jacobi-Davidson iteration method for linear eigenvalue problems. SIAM J. Matrix Anal. Appl. 17 (1996) 401–425.
Stewart G.W. and Sun Ji-Guang Matrix Perturbation Theory. Academic Press, San Diego,CA, 1990.
Tisseur F. and Dongarra J. A parallel divide and conquer algorithm for the symmetric eigenvalue problem on distributed memory architectures. SIAM J. Sci. Comput. 20 (1999) 2223–2236.
Trefethen L.N. and Bau D. III, Numerical Linear Algebra. SIAM, Philadelphia, PA, 1997.
Underwood R.R. An iterative block Lanczos method for the solution of large sparse symmetric eigenproblems. Technical Report STAN-CS-75–495,Department of Computer Science, Stanford University, Stanford, CA,1975.
van Kempen H.P.M. On quadratic convergence of the special cyclic Jacobi method. Numer. Math. 9 (1966) 19–22.
von Mises R. and Pollaczek-Geiringer H. Praktische Verfahren der Gleichungsaufl ősung. Z. Angew. Math. Mech. 9 (1929) 58–77,58–77
Watkins D.S. Some perspectives on the eigenvalue problem. SIAM Rev. 35 (1993) 430–471.
Watkins D.S. and Elsner L. Chasing algorithms for the eigenvalue problem. SIAM J. Matrix Anal. Appl. 12 (1991) 374–384.
Wilkinson J.H. The Algebraic Eigenvalue Problem. Clarendon Press,Ox ford, 1965.
Wilkinson J.H. and Reinsch C. (Eds.) Handbook for Automatic Computation. Vol. 2, Linear Algebra,S pringer,H eidelberg, 1971.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2003 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
He, W., Prabhu, N. (2003). Jacobi Decomposition and Eigenvalues of Symmetric Matrices. In: Aronov, B., Basu, S., Pach, J., Sharir, M. (eds) Discrete and Computational Geometry. Algorithms and Combinatorics, vol 25. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-55566-4_24
Download citation
DOI: https://doi.org/10.1007/978-3-642-55566-4_24
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-62442-1
Online ISBN: 978-3-642-55566-4
eBook Packages: Springer Book Archive