Abstract
The so-called Jacobi Procedure is extended to the case of normal matrices. A stable iterative procedure is described utilizing plane unitary transformations for such matrices which yield both the characteristic values and their associated vectors. Generally, the technique consists of minimizing at each stage the sum of the squares of the off-diagonal elements of the given matrix; however, there is one case in which this leads to no improvement; i.e. the lowest value for the change is non-negative. In this case, it is shown that a convergent procedure is still possible.
- 1 C. C J. JAcoBI, tVber ein leichtes Verf~hren, die in der Theorm der Sakularstorungen vorkommenden Gleiehungen numerisch aufzulosen, J Reme Angew. Math. 80, 51-95 (a846).Google Scholar
- 2 R. T. GRE~oRY, Computing eigenvalues and elgenvect, ors, Math. Tables A~ds Comp. 7, 215-220 (1953).Google Scholar
- 3 D A. POP:B AND C. TOMPKINS, Maximizing functions of rott~tions, J. Assoc. Comp. Math. 4;, 459-466 (1957) Google Scholar
- 4 G. E FORSV'rI~E AND P. HE,x'R~CJ, The cyclic J~cobI reef, hod for computing the principal values of a complex m~trix, to be published.Google Scholar
- 5 A. S. ~OUSEHOLDER, Pmnczptes of Numerzcal Analys~s, pp. 160-162, McGraw-Hill, New York, 1953Google Scholar
- 6 Zv~. LOTKIN, CharacLenstic val(ms of arbitrary mat, rices, Quarl Appl Math 14, 267~275 (1956).Google Scholar
- 7 J. (~REENSTADT, A method for finding roots of arbitrary matrices, Math Tables Aids Comp. 9, 47-52 (t955).Google ScholarCross Ref
- 8 l{. L. CAUSEY, Computing eigenvalues of non-Hermi~ian matrices by mebhods of Jacobi type, to be publishedGoogle Scholar
- 9 J yon Nsu>tA~'N, Mathematical Foundattons of Quantum Mechamcs, pp. 170-178 and pp. 223-229, Pr, nceton University Press, Princeton, 1955Google Scholar
- 10 J. H. M. W~DD~I~BURN, Lectures on Matrices, ch VII, American Mathematical Society Coil Publ 17, New York, 1934Google Scholar
- 11 H. g. GOLI)STINE, ~. J. MURRAY AND J. V(~N NEUMANN, The Jacobi method for real symmetric matrices, J. Assoc. Comp. Math. 6, 59-96 (1959). Google ScholarDigital Library
Index Terms
- A Procedure for the Diagonalization of Normal Matrices
Recommendations
The Arnoldi Method for Normal Matrices
For large Hermitian matrices the preconditioned conjugate gradient algorithm and the Lanczos algorithm are the most important methods for solving linear systems and for computing eigenvalues. There are various generalizations to the nonsymmetric case ...
Computing eigenvalues of normal matrices via complex symmetric matrices
AbstractComputing all eigenvalues of a modest size matrix typically proceeds in two phases. In the first phase, the matrix is transformed to a suitable condensed matrix format, sharing the eigenvalues, and in the second stage the eigenvalues ...
Normal Toeplitz Matrices
It is well known from the work of Brown and Halmos [J. Reine Angew. Math., 213 (1963/1964), pp. 89--102] that an infinite Toeplitz matrix is normal if and only if it is a rotation and translation of a Hermitian Toeplitz matrix. In the present ...
Comments