CVXOPT: A Python Package for Convex Optimization

At last, i can breath a sigh of relief. Seeing this python package made me realize how easy this project is gonna be. It's simple. I'll just make EMANS-OR dependent on CVXOPT.

From the site:

CVXOPT release 0.4 is a preliminary release of a Python package for convex optimization. It includes the following:

• Python classes for storing and manipulating dense and sparse matrices
• an interface to most of the double-precision real and complex BLAS
• an interface to the dense linear equation solvers and eigenvalue routines from LAPACK
• interfaces to the sparse LU and Cholesky solvers from UMFPACK and LDL
• routines for specifying linear programs and convex optimization problems with piecewise-linear objective and constraint functions
• a linear programming solver written in Python and interfaces to the LP solvers in GLPK and MOSEK.

Future releases will offer a more comprehensive set of convex optimization solvers (written in Python or based on external solvers) and modeling tools to support them.

The distribution also includes examples of convex optimization routines written in Python: an implementation of Newton's method for analytic centering, a maximum-likelihood covariance selection problem, an example illustrating norm and penalty approximation using linear programming, and a simple FIR filter design demo.

In developing CVXOPT we have several goals in mind:

• create a software environment that makes it straightforward to build applications that use convex optimization
• provide an interactive tool that allows students in convex optimization courses to easily specify and solve convex optimization problems
• facilitate implementation and 'rapid prototyping' of convex optimization algorithms.

The installation requires Python version 2.3 or higher and the following libraries (not included with the distribution): ATLAS or BLAS/LAPACK and UMFPACK.

CVXOPT is distributed under the GNU General Public License.