From help-octave-request at bevo dot che dot wisc dot edu Fri Dec 4 15:20:49 1998 Subject: Re: fminu-kind of function From: heberf To: jgomez at pc117179 dot shef dot ac dot uk Cc: help-octave at bevo dot che dot wisc dot edu Date: Fri, 4 Dec 1998 15:21:20 -0600 (CST) There is some contributed code at ftp.mathworks.com which does what you want. Look in pub/contrib/v4/optim/dsmax and there are three direct search algorithms for unconstrained problem. These maximize rather than minimize but you can just change the sign on your function to get the min (recall that minimizing a function is the same as maximizing the negative of the function). In fact one of the algorithms is the same one that is implemented in the Matlab function FMINS. Direct search algorithms are not the fastest but they are quite stable and don't get hung up at wierd points. I find these functions quite useful. I don't recall if you have to change anything to make them work in Octave. If so it's trivial. Heber > Hi!!! > I've been going through the mailing list's archive, and found someone > asked a while ago about fminu and fmins functions in Octave. At present, I have > been using fsolve, but the results I get seem to be random. Put it in other > words, I could do a better job myself :))) In particular, I just wanted to know > what's the best way to find the minimum of a function using octave. Is there > anything in fsolve that I should know about? At any case, my tries always get > fsolve to return a status of 3... > > Any help? > TIA, > Jose > -- > Jose L Gomez Dans PhD student > Radar & Communications Group > Department of Electronic Engineering > University of Sheffield UK > _______________________________________________________________________________ Heber Farnsworth Assistant Professor of Finance John M. Olin School of Business Washington University Campus Box 1133 phone: (314) 935-4221 One Brookings Drive FAX: (314) 935-6359 St. Louis, MO 63130-4899