GSoC 2015: Optimization Package: Non-linear and constrained least squares lsqcurvefit, lsqlin, lsqnonlin

[AsmaA]

Hi,

I am interested in the GSoC project for implementing Nonlinear and
constrained least squares optimization problem in Octave [1]. 

I am studying the Levenberg-Marquardt algorithm from [2]. I have taken a
look at the levmar.c library [3]. It provides wrapper mex files to call it
from MATLAB. The FAQ indicates the author heard that the wrappers are
already Octave compatible. But he isn't sure.

I have quite a bit of experience with MATLAB but not mex/Octave.

My next step is to learn mex, call levmar.c from MATLAB and run a few
examples between lsqcurvefit/lsqnonlin and levmar.c from MATLAB and
comparing data. Then call mex from Octave and do similar.

At this point, I'd like a little feedback if I'm proceeding in the right
direction. And any pointers would be appreciated :).

Other notes:
lsqnonlin in MATLAB also supports trust-region-reflective algorithm for
optimization. But I'll initially focus on levmar support.

Thank-you.

Kind Regards
Asma Afzal
PhD Student (Electrical Engineering) University of Leeds, UK.

References:

[1]
http://wiki.octave.org/Summer_of_Code_Project_Ideas#Nonlinear_and_constrained_least_squares

[2] http://www.imm.dtu.dk/pubdb/views/edoc_download.php/3215/pdf/imm3215.pdf

[3] http://users.ics.forth.gr/~lourakis/levmar/index.html




--
View this message in context: 
http://octave.1599824.n4.nabble.com/GSoC-2015-Optimization-Package-Non-linear-and-constrained-least-squares-lsqcurvefit-lsqlin-lsqnonlin-tp4668777.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.