On Tue, 27 Nov 2012, Rupert Swarbrick wrote:
> I haven't gone through your code carefully, but it just
> looks like your "dg" (gradient?) has gone to zero
... which I guess is good news, because it means the algorithm has
found the location of the extremum it's seeking to within machine
precision.
One potential source of confusion to watch out for: the same name
"method of steepest descents" is used for two very different
mathematical techniques. The other one, with which I'm more familiar,
is described at
<http://en.wikipedia.org/wiki/Method_of_steepest_descent>;.
--
HTH,
Dan