Hello,
I see that functions fMin and fMax can search for (constrained) extrema of functions of more than one variable (judging by a working example for fMin). This functionality is not documented. How exactly do fMin and fMax work in such cases?
Thanks.
fMin and fMax functions
Modérateur : xcasadmin
Re: fMin and fMax functions
It does use a public domain implementation of the cobyla algorithm.
Re: fMin and fMax functions
That's great, I wasn't aware that fMin and fMax are so powerful. I'll prepare some additions to the existing documentation so this functionality won't pass unnoticed by the future users. Also, a clone of Maple's NLPSolve is now just a few steps away, I'll post a new version of optimization.cc/.h, including nlpsolve function and the corresponding documentation over the weekend (which will together with lpsolve, minimize and maximize provide a full, free replacement of the Maple's Optimization package). Does the cobyla algorithm need a feasible initial point or that's not necessary?
Re: fMin and fMax functions
A guess is required, the syntax is fMin/fMax(expression,[constraints],[variables],[guess] with optional args: ...,eps,maxiter)
Re: fMin and fMax functions
Here's the updated documentation entry on fMin and fMax. I'm also posting a new version of optimization.cc/h in the respective thread.
- Pièces jointes
-
- fMin-fMax documentation.zip
- (1.02 Kio) Téléchargé 188 fois