Re: Optimization: Replace NaN by Inf

[Christoph Spiel]

On Mon, Aug 06, 2001 at 06:14:19PM +0100, Etienne Grossmann wrote:
>   when optimizing a function 'f', I am tempted to replace any NaNs
> that it returns by Infs. This way, the returned value can be
> compared w/ other values, and the optimization algorithm is happy.
> 
>   Can anyone think of a reason not to replace NaNs by Infs inside an
> optimization algorithm?

I would advise against replacing NaNs with Infs, especially if you are
using complex arithmetic.

Usually a NaN indicates that a computation has gone very wrong and the
cause should be investigated.  (This is one of the reasons why special
values like NaN have been introduced in the first place.)


-cls

-- 
Christoph L. Spiel  <address@hidden>
Hammersmith Consulting,  web: www.hammersmith-consulting.com
Phone: +49-8022-662908, fax: +49-8022-662909



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------