Groups | Search | Server Info | Login | Register

Groups > comp.soft-sys.math.maple > #179

Re: Numerical definition of variables

Show all headers | View raw

On 2011-06-14, Joe Riel <joer@san.rr.com> wrote:
> Herman Rubin <hrubin@skew.stat.purdue.edu> writes:

>> On 2011-06-13, rouben@shady.(none) (Rouben Rostamian) <rouben@shady> wrote:
>>> In article <slrnivcqre.jgv.hrubin@skew.stat.purdue.edu>,
>>> Herman Rubin  <hrubin@stat.purdue.edu> wrote:

>>>>As an example, suppose I have two equations r and rr, and
>>>>two variables b and c which occur in them.  Now when I enter

>>>>	fsolve({r,rr},{b,c})

>>>>it will produce the values of b and c.  What I want done is
>>>>to use this, without too much additional syntax, to DEFINE
>>>>the values of b and c for subsequent work.

>>> Will this do?

>>>    fsolve({r,rr},{b,c}); assign(%);

>> I tried this, and it gives the error message

>> Error (in assign) invalid arguments

> What was returned by fsolve?  It works for the simple case

> (**) eqs := { a + 3*b = 12,  b - a = 1 }:
> (**) fsolve(eqs, {a,b});                 
>                          {a = 2.250000000, b = 3.250000000}

> (**) assign(%);                          
> (**) a,b;
>                               2.250000000, 3.250000000


> My experience is that directly assigning to the solved variables
> is usually not the best way to work, at least not if you are
> want to deal with symbolic expressions.  The problem is that it
> then makes expressions that use those variables numeric, which
> frequently prevents further analysis.

> I'll generally assign the solution to a variable, say

>  fsol := fsolve( ... );

> and then, when I need numeric values for expressions, do

>   eval( some_expr, fsol );

What if the next thing you want to do is

fsolve(eq, c); 

where a and b are in eq?  Then a, b, and c are used in
some other calculations, including plotting.

I presume I could put all the conditions in one big
equation, but as the equations are necessarily 
ill-conditioned, will the computer manage to find
a close enough starting point?  This is why I want
to only have it do part, and use the plots to move
things around.


-- 
This address is for information only.  I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hrubin@stat.purdue.edu         Phone: (765)494-6054   FAX: (765)494-0558

Back to comp.soft-sys.math.maple | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

Numerical definition of variables Herman Rubin <hrubin@skew.stat.purdue.edu> - 2011-06-13 19:56 +0000
  Re: Numerical definition of variables rouben@shady.(none) (Rouben Rostamian) - 2011-06-13 20:17 +0000
    Re: Numerical definition of variables Herman Rubin <hrubin@skew.stat.purdue.edu> - 2011-06-14 00:04 +0000
      Re: Numerical definition of variables rouben@shadow.(none) (Rouben Rostamian) - 2011-06-14 01:04 +0000
      Re: Numerical definition of variables Joe Riel <joer@san.rr.com> - 2011-06-14 09:06 -0700
        Re: Numerical definition of variables Herman Rubin <hrubin@skew.stat.purdue.edu> - 2011-06-14 22:33 +0000
          Re: Numerical definition of variables Joe Riel <joer@san.rr.com> - 2011-06-14 16:24 -0700
      Re: Numerical definition of variables Peter Pein <petsie@dordos.net> - 2011-07-03 12:33 +0200

csiph-web