Re: exp(A) where A is a matix

[John W. Eaton]

On  4-Aug-1999, A. Scottedward Hodel <address@hidden> wrote:

| Haven't looked in the octave sources, but it would take less time to
| write
| 
|     #include <math.h>
|     ...
|     Matrix retval = ...;
| 
|     for (int ii = 0 ; ii < retval.rows() ; ii++)
|       for ( int jj = 0 ; jj < retval.columns() ; jj++)
|         retval(ii,jj) = exp(retval(ii,jj));
| 
| 
| than to dig through the code.  [Eventually, guilt would
| catch up to me and I'd have to search through the liboctave
| and src/DLD-FUNCTIONS sources to see how the exp function handles
| matrix arguments.]

I think you would find that it is implemented approximately the way
you have above, but it is done through a function pointer instead of
duplicating code for every function (sin, cos, tan, exp, log, etc.).

So, I don't think there is a

  Matrix exp (const Matrix& m)

function defined in the Octave sources.  But there are `map' and
`apply' functions defined for the matrix and vector classes.  They
allow you to write

  Matrix m, r;

  ...

  r = m.map (exp);

The `map' function is a const function, so it doesn't change the
object on which it operates.  The `apply' function is not a const
function and it modifies the object on which it operates.

jwe



---------------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.  To ensure
that development continues, see www.che.wisc.edu/octave/giftform.html
Instructions for unsubscribing: www.che.wisc.edu/octave/archive.html
---------------------------------------------------------------------