Efficient free / open source SOCP (second-order cone programming) solvers

Question

I am looking for a recommendation (or comparison) of solutions for programming a second-order cone with respect to the estimation speed. The solver must be free for non-commercial use or open source.

I am fairly frank about the environment: standalone solutions, libraries, Matlab, Python, R, etc. are acceptable.

My problem has a significant limitation in the constraints, which, I believe, can be used by good solvers to speed up the calculation.

Answer 1

you may want to take a look at the standard supported in

http://plato.la.asu.edu/bench.html

SOCP QP . , .

http://en.wikipedia.org/wiki/Second-order_cone_programming

, , .

, , MOSEK CPLEX, , , , , , .

, . .

-, , , . , . .

Ipopt - COIN-OR: Cbc:

, , , . , . , , mosek.com.

cbc: https://projects.coin-or.org/Cbc ipopt: https://projects.coin-or.org/Ipopt

, , , . , . , , mosek.com.

Answer 2

, , , cvxpy cvxopt ecos .
ecos , LP ( 3 , cvxopt ).
~ 5k c + python wrapper, ! scipy.sparse.csc; , .

Answer 3

CVXPY (http://www.cvxpy.org/), QCML (https://github.com/cvxgrp/qcml), C-, .

CVXPY is improving very fast. Problems in Check limitations in cvxpy with actual values ​​in order ... from a completely outdated version. Assuming your problem is not too large (less than a million variables), CVXPY is likely to satisfy your needs. Even with big problems, you can use the SCS solver in CVXPY to find a quick (albeit somewhat less accurate) solution.