On Mon, Jun 11, 2007 at 09:00:55AM -0600, Robert Dodier wrote:
> Hello,
>
> I have revised the find_root function so that it evaluates its arguments
> in the same manner as most other Maxima functions
...
> I didn't change the find_root algorithm; I wouldn't be
> surprised if stronger seach algorithms are known.
Robert, thank you very much for this improvement in find_root. It is
extremely valuable to be able to mix numerical and symbolic
computations in maxima, and I think this will significantly improve
the usability of find_root.
I have been relatively satisfied with the performance of find_root,
but I haven't exercised it heavily (for example, in a tight loop, or
doing the numerical integral of a function defined by find_root
etc). Brent's method is a popular but slightly complicated method. If
speed becomes an issue, perhaps we may consider modifying the
algorithm. For the moment I think it is reliable and acceptably fast
for me at least.
--
Daniel Lakeland
dlakelan at street-artists.org
http://www.street-artists.org/~dlakelan