Does the linearalgebra package have switches to allow the calculations
to be done with arbitrarily large precision?
It has been said on this list that some numerical routines have
limited precision because some of the lisp programs used have this
limitation built in.
I think one should work toward having the ability to put in large
precision in any routine that maxima does.
By this, I do not mean pure symbolic calculations. In some
situations that is either impossible or impractical.
-sen
On Tue, 17 Jul 2007, Andrej Vodopivec wrote:
> Hi,
>
> in current cvs mnewton is broken. mnewton calls linsolve with
> keepfloat:true. This produces errors (from bugreport 1725549):
>
> (%i1) load(mnewton);
> (%o1) /usr/local/share/maxima/5.12.0cvs/share/contrib/mnewton.mac
> (%i2) trace(linsolve)$
> (%i3) mnewton([x1+3*log(x1)-x2^2, 2*x1^2-x1*x2-5*x1+1], [x1, x2], [5, 5]);
>
> < some lines removed >
>
> [[-5.6150895140665*h[2]+1.784746613146*h[1]-.036397473842489,-
> 3.8228900255754*h[2]+7.4840153452685*h[1]+0.3815913193922],[h[1],h[2]]]
> 1 Exit linsolve [h[1]=0,h[2]=0] <---- WRONG!
> (%o3) [[x1=3.8228900255754,x2=2.8075447570332]]
>
> Since I don't think linsolve is intended to solve equations with
> floating point coefficients, I propose we remove the call to linsolve
> and use the linearalgebra package. A patch to do this is attached. The
> patch also fixes bug 1666011.
>
>
--
---------------------------------------------------------------------------
| Sheldon E. Newhouse | e-mail: sen1 at math.msu.edu |
| Mathematics Department | |
| Michigan State University | telephone: 517-355-9684 |
| E. Lansing, MI 48824-1027 USA | FAX: 517-432-1562 |
---------------------------------------------------------------------------