On Nov 19, 2007 9:24 AM, Rogers, Raymond <raymond-rogers at idexx.com> wrote:
> You might try implementing Newton-Raphson....
> <http://numericalmethods.eng.usf.edu/ebooks/newtonraphson_03nle_ebook.htm>I'm
> really surprised that Maxima doesn't have a package to do this.
Maxima has a variety of documented numerical solving routines. If you're
looking for Newton method, try
?? newton
at the command line, which documents naive univariate Newton-Raphson (newton)
and multivariate Newton-Raphson (mnewton). The query
?? root
will show find_root, a combination of binary search and Newton-Raphson which
works for a larger variety of functions, realroots, which uses Sturm
sequences for univariate polynomials (real_root), allroots for complex roots
of polynomials. And so on.
Why doesn't solve use any of these methods? Because Maxima is primarily a
*symbolic* system, so solve looks for closed-form symbolic solutions.
-s