Summary: algorithm for eigenvalues



 Dear CCL-ers,
 Last week I posted a question about eigenvalaues of a large
 banded hermitian matrices. I got several responses which I
 summarized here below.
 The best way to do this is to use the Householder reduction and
 after that an QR method. With all the answers I recieved,I will
 try to modify the algorithm so it would fit strictly to my problem.
 For theoretical explenation:
 Wilkinson, "The Algebraic Eigenvalue Problem", Clarendon Press, Oxford
 (1965)
     (there should be a newer edition, but I didn't find it in our library)
 Press et. al., Numerical recipes in C, Cambridge University Press",
 Cambridge
  (1992)
     (it's now also on the web: http://cftata2.harvard.edu/nr)
 Burden and Faires, "Numerical Analysis", PWS Publishing Company,
 Boston
  (5th ed., 1993)
     (This is a , IMHO, good introductionary book in numerical analysis,
      It could be used in such a course for computational chemists.)
 Wilkinson and Reinsch, "Handbook for Automatic Computation, Vol. II Linear
   Algebra", Springer-Verlag, Berlin (1971)
 Many have already implemented the algorithms. They are stored in several
 packages.
 Here are some:
 LAPACK: http://www.netlib.org/lapack
 EISPACK: http://www.netlib.org/eispack
 ARPACK: http://www.caam.rice.edu/~kristyn/parpack_home.html
 Thanks to all who responded,
 Niels
 +----------------------------------------------------------+
 | Niels J. van der Laag                                    |
 |      Student Computational Chemistry                     |
 |        Dept. of Physical Chemistry  (Solid State NMR)    |
 |        University of Nijmegen                            |
 |        Toernooiveld 1, NL-6525 ED Nijmegen, Netherlands  |
 |        phone: ++31-24-3653112 email: nila' at \`solidmr.kun.nl |
 |                                      nielsl' at \`sci.kun.nl   |
 +----------------------------------------------------------+