[Date Index] [Thread Index] [Author Index]

Re: FindMinimum Termination Criteria?

• To: mathgroup at smc.vnet.net
• Subject: [mg4787] Re: FindMinimum Termination Criteria?
• From: rubin at msu.edu (Paul A. Rubin)
• Date: Mon, 16 Sep 1996 23:51:16 -0400
• Organization: Michigan State University
• Sender: owner-wri-mathgroup at wolfram.com

In article <51c50q$b2u at ralph.vnet.net>,
   dreece at atl.mindspring.com (Daryl Reece) wrote:
->Could someone tell me what criteria is used in FindMinimum to find the
->minimum point and to terminate the search?  I'm running FindMinimum on
->a function that does not have local minima yet it terminates.  I've
->tried upping the MaxIterations and the AccuracyGoal but the results
->don't seem to improve.
->
->Thanks in advance,
->Daryl

It obviously can stop if it hits an iteration limit, but you say boosting 
the limit doesn't help.  It will stop if you give it a domain interval for 
one or more of the variables and it finds itself trying to leave that 
interval, but presumably that's not it (you'd know if that was happening). 
 Otherwise, I think it stops because the derivative/gradient is looking too 
much like a zero.  Have you calculated the derivative (gradient?) at the 
"solution" it gives?  Could it be, if not a critical point, at least close 
to one?  (Also, keep in mind that if it ever stumbles onto a local max, the 
zero derivative will stop it there as well.)

Come to think of it, maybe I should ask first if the function *has* a 
derivative everywhere?

-- Paul

**************************************************************************
* Paul A. Rubin                                  Phone: (517) 432-3509   *
* Department of Management                       Fax:   (517) 432-1111   *
* Eli Broad Graduate School of Management        Net:   RUBIN at MSU.EDU    *
* Michigan State University                                              *
* East Lansing, MI  48824-1122  (USA)                                    *
**************************************************************************
Mathematicians are like Frenchmen:  whenever you say something to them,
they translate it into their own language, and at once it is something
entirely different.                                    J. W. v. GOETHE

==== [MESSAGE SEPARATOR] ====