levmar  2.6 Description:

levmar for Linux - levmar is an implementation of the Levenberg-Marquardt nonlinear least squares algorithms in C/C++.

levmar is an implementation of the Levenberg-Marquardt nonlinear least squares algorithms in C/C++. The lmder routine from Minpack, implemented in the early '80s at the Argonne National Lab, is perhaps the most widely used free implementation of the LM algorithm. lmder is written in FORTRAN77 and over the years has proven to be a reliable piece of software. Considering that FORTRAN routines can be called from C/C++, one might wonder about the motivation for writing a version of LM in C. Well, the problem is that when FORTRAN is called from C, the programmer should be aware of (and conform to) several rules regarding name mangling, argument passing, multidimensional array memory layout, linkage conventions, etc, that are unnatural compared to ordinary C rules. A second reason is that this approach takes for granted that a FORTRAN compiler for the target programming environment is available, which might not necessarily be the case. Another reason has to do with failure to understand the inner workings of a FORTRAN implementation: Occasionally, when it is necessary to precisely understand what the FORTRAN code does, certain pieces of it might seem incomprehensible to programmers without any knowledge of FORTRAN. Automatic FORTRAN to C translators (e.g. f2c) do not solve the problem since the produced C code is pretty illegible to "uninitiated" humans. Moreover, documentation describing the mathematics upon which the implementation is based might be obscure or inaccessible. Last but not least, a candidate LM implementation in C should be free and technically sound. For example, the C variant of the LM algorithm presented in the "Numerical Recipes" book (i.e. mrqmin), is not always a viable choice: Besides its being copyrighted, it is reputed to lack robustness. For the above reasons, I have developed the levmar package which includes C implementations of LM flavors that are also usable with C++. levmar includes double and single precision LM implementations, both with analytic and finite difference approximated Jacobians. It is provided free of charge, under the terms of the GNU General Public License. The mathematical theory behind unconstrained levmar is described in detail in the lecture notes entitled Methods for Non-Linear Least Squares Problems, by K. Madsen, H.B. Nielsen and O. Tingleff, Technical University of Denmark; Matlab implementations of the algorithms presented in the lecture notes are also available. Note however that the formulation of the minimization problem adopted here is slightly different from that described in the lecture notes.


To conclude levmar works on Linux operating system and can be easily downloaded using the below download link according to GNU Public License license. The download file is only 0.08 MB  in size.
levmar was filed under the Engineering category and was reviewed in softlookup.com and receive 2.8/5 Score.
levmar has been tested by our team against viruses, spyware, adware, trojan, backdoors and was found to be 100% clean. We will recheck levmar when updated to assure that it remains clean.

levmar user Review

Please review levmar application and submit your comments below. We will collect all comments in an effort to determine whether the levmar software is reliable, perform as expected and deliver the promised features and functionalities.

Popularity 5.6/10 - Downloads - 71 - Score - 2.8/5

Category: Engineering 
Publisher: Manolis Lourakis
Last Updated: 01/11/2019
Requirements: Not specified
License: GNU Public License
Operating system: Linux
Hits: 284
File size: 0.08 MB 
Price: Not specified

Leave A comment
Name: *
E-Mail: *
Comment: *