[Octave-bug-tracker] [bug #30092] eig/eigs: possibly incorrect result of eigenvalue matrix

[anonymous]

URL:
  <http://savannah.gnu.org/bugs/?30092>

                 Summary: eig/eigs: possibly incorrect result of eigenvalue
matrix
                 Project: GNU Octave
            Submitted by: None
            Submitted on: Wed 09 Jun 2010 09:11:39 AM UTC
                Category: None
                Severity: 3 - Normal
              Item Group: Incorrect Result
                  Status: None
             Assigned to: None
         Originator Name: massimiliano culpo
        Originator Email: address@hidden
             Open/Closed: Open
         Discussion Lock: Any
                 Release: 3.3.51
        Operating System: GNU/Linux

    _______________________________________________________

Details:

The following code produces a wrong result:

octave:1> B = [1,2; 0 4]
B =

   1   2
   0   4

octave:2> [V,D] = eig(B)
V =

   1.00000   0.55470
   0.00000   0.83205

D =

Diagonal Matrix

   1   0
   0   4

octave:3> V'*V
ans =

   1.00000   0.55470
   0.55470   1.00000

The matrix V is not unitary at all. I wonder if anyone can reproduce this bug
, or if it is a problem of my installation of octave.

This happens with octave 3.2.4 and 3.3.51+ compiled on Ubuntu 9.10. I don't
know if the information can be useful, but I linked against qrupdate at
compile time.

The same happens with eigs. Try for instance with the matrix:

octave:5> C = kron(eye(3,3),B)
C =

   1   2   0   0   0   0
   0   4   0   0   0   0
   0   0   1   2   0   0
   0   0   0   4   0   0
   0   0   0   0   1   2
   0   0   0   0   0   4

M.




    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?30092>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/