| [eigen] Levenberg Marquardt parameters |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: [eigen] Levenberg Marquardt parameters
- From: Koldo Ramirez <koldo.ramirez@xxxxxxxxx>
- Date: Sun, 19 Jun 2011 23:46:32 +0200
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:date:message-id:subject:from:to :content-type; bh=9AYUIW7EKGemINGl3IdE4/1yucYWW/eSv2M/0e3/ncE=; b=LBeBwvlkIDohv/yMAT9RrlP0J4f4cHO1XqSlnVqgTEPbyY1a+p+di4SSO1REHPXprK vWyjeVTbxdW8Vw0JYnI6dkCl9SH0Gu0tcsDmtITYRPyBsZHTKfb5TnRYU2ozc6sULkoP ziEk/3iDs2twWJpSE+D+E0whUhV5QL2jWLLjk=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=idv7en2/yOmTC5QmP5/2NbcUx3OC6XlJbqCs4NaprTICcFyeQfMqmXZJ3Sg00Xc0BG 4j4CI1PY/on4Cxpm6F3VSTlvEM5xZxxd3KswtFlMm3/F9E/jL+LbE/rasNfuFGFel7n5 Kco7tiKUn6RYBOHMnPWKSR24vAcOIrmOxIG7s=
Hello Thomas
I am really enjoying Non Linear Optimization module. It lets me solve very
complex things in a smart way :-) .
However I have had a hard time this weekend about a 17 unknowns set of
equations minimized with Levenberg Marquardt.
With initial values very close to results it worked well, but putting
realistic initial values the system did not converge. It returned a
TooManyFunctionEvaluation (5),
Checking the code, it seems that in the constructor, the max number of
function evaluations (fev) is:
maxfev (400)
But old lmder1() function, before calling minimize(), inits some parameters:
n = x.size();
....
parameters.maxfev = 100*(n+1);
So in my case, maxfev == 400 instead of maxfev == 100*(17+1) == 1800.
Doing this change, the system converged with good results after nfev ==
1619..., What a relief !
In
http://eigen.tuxfamily.org/dox-devel/unsupported/group__NonLinearOptimization__Module.htmlyou
state that all minpack methods like lmder1() are deprecated.