Visit www.embotech.com/ECOS for detailed information on ECOS.
ECOS is a numerical software for solving convex second-order cone programs (SOCPs) of type
min c'*x
s.t. A*x = b
G*x <=_K h
where the last inequality is generalized, i.e. h - G*x
belongs to the cone K
.
ECOS supports the positive orthant R_+
, second-order cones Q_n
defined as
Q_n = { (t,x) | t >= || x ||_2 }
with t a scalar and x
in R_{n-1}
,
and the exponential cone K_e
defined as
K_e = closure{(x,y,z) | exp(x/z) <= y/z, z>0}
where (x,y,z)
is in R_3
.
The cone K
is therefore
a direct product of the positive orthant, second-order, and exponential cones:
K = R_+ x Q_n1 x ... x Q_nN x K_e x ... x K_e
Through a recent extension by Han Wang, ECOS now comes with a branch-and-bound procedure (a direct translation of Stephen Boyd's lecture slides) called ECOS_BB
for solving mixed-integer or mixed-boolean programs of the form
min c'*x
s.t. A*x = b
G*x <=_K h
some x_i in {0,1}
some x_j integer
Note: the branch-and-bound module has been designed to solve small problems at acceptable speeds and with minimum added code complexity (ca. 200 lines of code on top of ECOS).
ECOS has numerous interfaces, each hosted in a separate git repository. The core ECOS solver (this repository) is included in the interface repositories as a submodule. You should run git submodule init
and git submodule update
after cloning the interface repositories.
Please refer to the corresponding repositories or the embotech documentation for information on how to install and use the software.
ECOS is distributed under the GNU General Public License v3.0. Other licenses for the core solver may be available upon request from embotech.
A full list of features, documentation and installation instructions as well as pre-compiled Matlab MEX binaries can be found here.
The solver is essentially based on Lieven Vandenberghe's CVXOPT ConeLP solver, although it differs in the particular way the linear systems are treated.
The following people have been, and are, involved in the development and maintenance of ECOS:
- Alexander Domahidi (principal developer)
- Eric Chu (Python interface, unit tests)
- Stephen Boyd (methods and maths)
- Michael Grant (CVX interface)
- Johan Löfberg (YALMIP interface)
- João Felipe Santos, Iain Dunning (Julia inteface)
- Han Wang (ECOS branch and bound wrapper)
- Santiago Akle (Exponential cone method and extension)
The main technical idea behind ECOS is described in a short paper. More details are given in Alexander Domahidi's PhD Thesis in Chapter 9.
The algorithm for exponential cones can be found in Santiago Akle's PhD Thesis in Chapter 10.
If you find ECOS useful, you can cite it using the following BibTex entry:
@INPROCEEDINGS{bib:Domahidi2013ecos,
author={Domahidi, A. and Chu, E. and Boyd, S.},
booktitle={European Control Conference (ECC)},
title={{ECOS}: {A}n {SOCP} solver for embedded systems},
year={2013},
pages={3071-3076}
}