The LA_TRIQL procedure uses the QL and QR variants of the implicitly-shifted QR algorithm to compute the eigenvalues and eigenvectors of a symmetric tridiagonal array. The LA_TRIRED routine can be used to reduce a real symmetric (or complex Hermitian) array to tridiagonal form suitable for input to this procedure.
LA_TRIQL is based on the following LAPACK routines:
The following example computes the eigenvalues and eigenvectors of a given symmetric array:
; Create a symmetric random array:
n = 4
seed = 12321
Array = RANDOMN(seed, n, n)
array = array + TRANSPOSE(array)
; Reduce to tridiagonal form
q = array ; make a copy
LA_TRIRED, q, d, e
; Compute eigenvalues and eigenvectors
eigenvalues = d
eigenvectors = q
LA_TRIQL, eigenvalues, e, eigenvectors
PRINT, 'LA_TRIQL eigenvalues:'
-3.90712 -2.53874 2.14756 4.41393
A named vector of length n containing the real diagonal elements, optionally created by the LA_TRIRED procedure. Upon output, D is replaced by a real vector of length n containing the eigenvalues.
The (n - 1) real subdiagonal elements, optionally created by the LA_TRIRED procedure. On output, the values within E are destroyed.
An optional named variable that returns the eigenvectors as a set of n row vectors. If the eigenvectors of a tridiagonal array are desired, A should be input as an identity array. If the eigenvectors of an array that has been reduced by LA_TRIRED are desired, A should be input as the Array output from LA_TRIRED. If A is not input, then eigenvectors are not computed. A may be either real or complex.
Set this keyword to use double-precision for computations and to return a double-precision (real or complex) result. Set DOUBLE = 0 to use single-precision for computations and to return a single-precision (real or complex) result. The default is DOUBLE = 0 if none of the inputs are double precision. If A is not input, then the default is /DOUBLE if D is double precision. If A is input, then the default is /DOUBLE if A is double precision (real or complex).
Set this keyword to a named variable that will contain the status of the computation. Possible values are:
- STATUS = 0: The computation was successful.
- STATUS > 0: The algorithm failed to find all eigenvalues in 30n iterations. The STATUS value specifies how many elements of E have not converged to zero.
Note: If STATUS is not specified, any error messages will be output to the screen.
Resources and References
For details see Anderson et al., LAPACK Users' Guide, 3rd ed., SIAM, 1999.