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

Re: Findminimum Question

• To: mathgroup at smc.vnet.net
• Subject: [mg24298] Re: Findminimum Question
• From: Jens-Peer Kuska <kuska at informatik.uni-leipzig.de>
• Date: Fri, 7 Jul 2000 00:11:38 -0400 (EDT)
• Organization: Universitaet Leipzig
• References: <8k0ups$q6k@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

Hi,

The appendix of The Mathematica Book A 9.4
say

With Method->Automatic, FindMinimum uses various methods due to Brent:
the conjugate gradient in one dimension, and a modification of Powell's
method in several dimensions. 

If the function to be minimized is a sum of squares, FindMinimum uses
the Levenberg-Marquardt method (Method->LevenbergMarquardt). 

With Method->Newton FindMinimum uses Newton's method. With
Method->QuasiNewton FindMinimum uses the BFGS version of the
quasi-Newton method.

But you may have a look into the Statistics`NonlinearFit` standard
package.
It can calculate the errors for you.

Regards
  Jens

ayyer at my-deja.com wrote:
> 
> Hi,
>         I was wondering if I could know the principle on which Mathematica uses
> the FindMinimum function. I mean, does it use a common algorithm ( like
> the Golden Search Algorithm) for all functions or does it analyse the
> functions by some of their specific properties, like continuity,
> differentiability, or shallow/deep regions and the like.
>         I needed to know this to find out the error bars for the function i am
> using. In case it is important, i am just minimising a sum of squares of
> a function to find the best fit.
> 
> Thanks in advance,
> Ayyer
> 
> Sent via Deja.com http://www.deja.com/
> Before you buy.