Subject: denominator of really complicated expression
From: Ryan Krauss
Date: Wed, 23 Nov 2005 16:27:25 -0500
It seems like with grind:true, stringout is almost as fast as save and
then the parsing problem is much less challenging (just get rid of the
indentation and the newlines I think).
Thanks Robert! That saved me a lot of work.
Ryan
On 11/23/05, Robert Dodier wrote:
> On 11/23/05, Ryan Krauss wrote:
>
> > So, Robert was very much correct about the speed of save vs. stringout:
> > save takes 0.14 seconds stringout takes roughly 300 seconds
> >
> > So that's about a 2000 times speed up.
> >
> > But, this leaves me with a bit of a daunting parsing task.
>
> I'm sorry to send mixed signals, but on second thought,
> the time spent attempting to parse the "save" output is
> likely to be much greater than the time spent on "stringout" ...
>
> As it happens there are a couple of other ways to get
> printed output. Trying out the ones I can think of,
> it seems that setting grind : true$ and then calling
> stringout is faster than stringout with
> grind = false (the default). With grind = true, stringout
> doesn't try to put everything on one line; hopefully
> it's easy enough to join the lines together to make
> the Python parser happy.
>
> About Maxima functions for optimization.
> There is an implementation of the simplex method for
> linear programming (not Nelder-Mead) and a couple of
> implementations of the augmented Lagrangian method for
> constrained nonlinear optimization. But there isn't a
> conjugate gradient, quasi-Newton, or Nelder-Mead,
> so far as I know.
>
> Hope this helps,
> Robert Dodier
>