ok -- I've tried the following:
?
I replaced
?
'integrate(taylor(...
?
?
with
?
'integrate('taylor(...
?
to avoid that taylor tries to work on symbolic expressions, since
most of the variables are explained only later.
I've used
?
compile_file(...);
load(second(%));
?
to exploit compilation...
?
I've installed the latest version of maxima (5.29.1).
?
Never mind -- I'm debugging...
After one unsuccessful day yesterday I've only tried to exploit some more trained eyes....
?
?
Peter
--- On Sun, 12/16/12, Stavros Macrakis <macrakis at alum.mit.edu> wrote:
From: Stavros Macrakis <macrakis at alum.mit.edu>
Subject: Re: [Maxima] taylor series
To: foelsche at sbcglobal.net
Cc: maxima at math.utexas.edu, "Robert Dodier" <robert.dodier at gmail.com>
Date: Sunday, December 16, 2012, 2:17 PM
Peter,
I agree completely with Robert Dodier's "PS" -- we can't really give you useful advice without that background. ?Even if we manage to help you without that, I find it presumptuous on your part to just dump a source file in our laps and say "fix my problem".
For example, it may be that you can solve your problem by calculating with truncated Taylor series, something like this:
? ? ? ? term1: taylor(expr1,x,0,5)$
? ? ? ? term2: taylor(expr2,x,0,5)$
? ? ? ? term1*term2;
instead of?
? ? ? ? taylor(expr1*expr2,x,0,5)$
It may be that substituting numerical values for your parameters earlier in the process makes sense.
It may be that you should really be using numerical techniques and not symbolic for the whole calculation.
It may be that Robert's suggestion of calculating with a formal Taylor series is best.
etc.
I really don't know, and?I'm not going to try to reverse-engineer your code to see if some other approach is more appropriate.
? ? ? ? ? ? ? ? -s
On Sun, Dec 16, 2012 at 4:58 PM, Robert Dodier <robert.dodier at gmail.com> wrote:
On 2012-12-16, =?utf-8?Q?foelsche at sbcglobal.net?= <foelsche at sbcglobal.net> wrote:
> Can anybody tell me, why this maxima source is not finishing....
Well, the problem is that the expressions you are working with are
enormous -- one of them has 100,000 terms or something like that. Maxima
is doing its best but there may not be enough time & memory to finish
the job.
As to the reason why the expressions are so big -- it's not clear if
it's inherent in the problem, or perhaps Maxima's heuristics for
simplification are actually working towards bigger & bigger expressions,
or something else.
I am working on an approach which keeps the expressions relatively
small. Basic idea is to construct a formal Taylor series & integrate it,
then substitute the actual expressions & actual values later. Trying to
construct the Taylor series seems to massively expand an already-large
expression ... I'll get back to you later today or tomorrow about it.
best
Robert Dodier
PS. Heroic debugging efforts notwithstanding, it would really help if
you could indicate what you have tried, what worked, what didn't, what
you're trying to accomplish, etc.
_______________________________________________
Maxima mailing list
Maxima at math.utexas.edu
http://www.math.utexas.edu/mailman/listinfo/maxima