Welcome to the OctMOSEK project!

Purpose

The Octave-to-MOSEK Optimization Interface is designed to make the optimization facilities of MOSEK available from Octave. MOSEK is a software library able to solve large-scale optimization problems, and supports:

Linear programming (LP)
Quadratic and All-Quadratic programming (QP and QCQP)
Second-Order Cone Programming (SOCP)
And all the above with Mixed-Integer variables..

Note that the MOSEK libraries ships with trial and academic licenses free of charge, but is otherwise a proprietary product.

System Requirements

You need Octave! Obviously, since this is an interface for Octave. However, since stable 64 bit versions does not yet exist for all platforms you might be limited to 32 bit Octave linking to 32 bit MOSEK.

You need MOSEK! Again obviously, as this is an interface to MOSEK. However, only MOSEK version 6.0.0.122 and newer are supported. Free trial and academic licenses are available, and quotations can be requested to allow commercial usage.

You need the OctMOSEK source code! We do not distribute binary versions of the OctMOSEK package. Instead, you are referred to the *.tar.gz file under the "Downloads" tab on this website.

Installation

The OctMOSEK package features auto-configuration which will work in most cases if the system has been properly set up (see System Requirements). Inserting the correct location and version below, most users will be able to install the package with:

pkg install -verbose 'LOCATION/octmosek_VERSION.tar.gz'

If this does not work, we advice you to look in the userguide for more information.

Getting started

The package can be loaded with

pkg load octmosek

The provided functions can be seen with

pkg describe -verbose octmosek

The manual for each function (here mosek) can be found with

help mosek

The mosek function is used to solve all supported optimization problems.

Example 1
In this example a linear program is defined and solved.

clear -v lo1;
lo1.sense = "max";
lo1.c = [3 1 5 1];
lo1.A = sparse([3 1 2 0;
                2 1 3 1;
                0 2 0 3]);
lo1.blc = [30 15 -Inf];
lo1.buc = [30 inf 25];
lo1.blx = [0 0 0 0];
lo1.bux = [Inf 10 Inf Inf];
r = mosek(lo1);

Example 2
In this example the formulation of a mixed integer linear program is read from a file and solved. The first line finds the root of the OctMOSEK package, and the second extracts the path of the milo1.opf file within. Alternatives are lo1.opf and cqo1.opf.

pkg_root = pkg("list"){strcmp({[pkg("list"){:}].name},"octmosek")}.dir;
modelfile = fullfile(pkg_root, "extdata", "milo1.opf");
r_read = mosek_read(modelfile);
r_solve = mosek(r_read.prob);

Tips and tricks, questions, feature requests and bug reports

We invite all users of the Octave-to-MOSEK Interface to join our central MOSEK discussion group. This group is particularly intended for academics using MOSEK in their research.