MathGroup Archive 1997

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

Search the Archive

Re: FindMinimum

  • To: mathgroup at smc.vnet.net
  • Subject: [mg6990] Re: FindMinimum
  • From: daiyanh at mindspring.com (Daitaro Hagihara)
  • Date: Thu, 1 May 1997 14:48:37 -0400 (EDT)
  • Organization: MindSpring Enterprises
  • Sender: owner-wri-mathgroup at wolfram.com

In article <5j758o$5nu at smc.vnet.net>, Sean Ross <SEAN at mail.creol.ucf.edu> wrote:

> observation:  FindMinimum rarely changes more than one parameter at a 
time.  It
> usually minimizes first changing only x, then changing only y, then 
changing only
> z etc.  Only occasionally will it make a guess by changing two parameters.

My usual amateuristic comments...
You probably have stiff set of equations.  That's the reason why only one 
coordinate with the largest derivative is changing at a time.  This is 
readily seen in Davidon-Fletcher-Powell method (as in NR).  I personally 
don't know any algorithm out there that does what you describe (but I'm no 
pro in this either).

Daitaro Hagihara


  • Prev by Date: Nonlinear fitting with integral?
  • Next by Date: Re: Writing Mathematica Plots to Files as PICTS?
  • Previous by thread: Nonlinear fitting with integral?
  • Next by thread: Re: Writing Mathematica Plots to Files as PICTS?