Posted by Terry J. Hilsabeck on August 02, 1998 at 01:30:16:

In Reply to: eigenvalue problem posted by K.A.Hafez on May 01, 1998 at 02:44:02:

I use the lapack function dgeev for eigenvalues of

a real non-symmetric matrix and it works very well.

: My Dear:

: I encounter a very strange/severe problem concerning the determination of the eigenvalue and the associated eigenvectors. I know eigenvalue problem is a very elementary problem nowadays, but really, I coldn't solve it myself. Even, I wrote a program for this, it didn't give rasonable results

: My problem concerning the solution of the following eigenvalue problem:

: {[K]-c^2[M]}.X=0.0

: where:

: K : is smmetric band real stiffness matrix (154x154)

: M: is antisymmetric real mass matrix (154x154)

: c: is the natura frequency or the eigenvalues

: X: is the eigenvector

: I already tried some routines called RGG and RG from the netlib, but any of them didn't give correct results. In fact all of them gave very strange results.

: Even I tried some checks to check the algorithm which already incorporated inside the subroutine, but I couldn't decide its suitablitiy.

: I would appreciate it if you are kind enoug and try to introduce any kind of help as soon as you can.

: waiting in front of the screen.

: sincerely,

: K.A.Hafez