Implementation of Levmar 2.6 Brafange optimization algorithm

Levmar is a http://www.aliyun.com/zixun/aggregation/17547.html "> c++/c++ implementation of a powerful and efficient LM Lafange optimization algorithm. LM solves the problem of nonlinear least squares and is equipped with an M-observation model, which is a nonlinear unknown parameter (m>=n). The Levmar includes a double and a single precision LM variable, an analysis and a finite difference Jacobian matrix. It also supports some constrained nonlinear least squares, and can use linear equations and box restrictions.

Levmar 2.6 This version adds the diagonal scaling xlevmar_bc_der () function, a linear system solver that supports parallel Cholesky decomposition with plasma, as a linear algebraic library of multiple core processors. The problem of linear solver has been fixed so that they are in symmetric matrices, thus better caching performance for lower triangular operations. The CMake configuration file for the construction project was revised.

Software Information: http://www.ics.forth.gr/~lourakis/levmar/

Download Address: http://www.ics.forth.gr/~lourakis/levmar/levmar-2.6.tgz