On 4/14/07, Albrecht Frenzel <afrenzel at 01019freenet.de> wrote:
> LG(x,p):=Round(log(x)*p)/p;
> Round(v):=floor(v+0.5);
> Td(t,h):=-((1188500*t+282506450)*log(h)+20301957*t)/((5000*t+1188500)*log(h)-20301957);
> TD(t,h,p):=-Round(((1188500*t+282506450)*LG(h,p)+20301957*t)/(Round(((5000*t+1188500)*LG(h,p)-20301957)/P)));
> P:1000;
> E(t,h,p):=(Td(t,h)*P-TD(t,h,p))/P;
> Next I would like to compute min and max on some 2D interval
> instead of plot3d.
Albrecht, I don't think Maxima has the capability to solve problems
such as this symbolically. Numerically, there is a function lbfgs to
minimize a function making use of the gradient of the function.
If the gradient is not available, I don't think Maxima has anything at
present. This web page
http://www-swiss.ai.mit.edu/~gjs/6946/
has a link to some code which includes Brent's method, in Scheme
but it would be straightforward to adapt to Lisp. (The code is GPL so
it seems like we might see what else could be adapted for Maxima.)
Looking at the plot made by plot3d for the function above, it is very
bumpy. So it could well be that it is going to be difficult for any
numerical method.
Hope this helps,
Robert Dodier