In this chapter we cover the basics of starting to use RepLAB.

Downloading the library

The representation theory code of the library is self-contained. Extended features such as unit tests, code coverage and convex optimization make use of external code. Here are two ways of installing the library with the desired set of features to get started. Choose the one which suits you best.

Option 1: Download the latest RepLAB release, and use our easy install script.

  1. Download the latest release ZIP, see the GitHub release page for all releases. This ZIP file includes the core code, but not the external dependencies.

  2. Launch MATLAB/Octave, run the replab_init autoinstall script in the root folder which will take care of downloading and installing dependencies, which are:

  • to run tests: MOxUnit

  • to define convex optimization (SDP) problems and run corresponding tests: YALMIP

  • to solve SDP problems and run corresponding tests: SDPT3

  1. Go to the “Initializing the library” section below and follow the usage instructions there.

Option 2: Clone the library

Initializing the library

To use the library, the RepLAB folder must be added in Matlab or Octave. Additional paths are also necessary to enable specific functionalities, as mentioned above, and a few variables mush be initialized. This can be done with


which should lead to the following output

>> replab_init
Adding RepLAB to the path
Initializing dependency vpi
Initializing dependency YALMIP
Initializing dependency sdpt3
Adding embedded SDPT3 solver to the path
Initializing dependency MOxUnit

This command checks in particular whether an instance of YALMIP is available and configured on your system. If this is not the case, the embedded version of yalmip is used. RepLAB uses the YALMIP interface to solve convex optimization problems. The replab_init command also ensures that an SDP solver such as SeDuMi is properly set up. If this is not the case, it activates the embedded SDPT3 solver. The proper installation of a YALMIP instance can be checked with the command yalmiptest.

The command replab_init should always be used before running any RepLAB command. This command only takes some time to run the first time it is called in a MATLAB/Octave session.

Customizing the initialization script

The script replab_config.m present in the root folder can be customized: for example, unneeded dependencies can be commented out, and various RepLAB options can be configured.

This script is run once during the replab_init initialization; subsequent calls to replab_init will not run it again, except if a clear all command has been run in between.


The proper installation of RepLAB can then be checked by running the test commands:


This checks the proper working of the whole package (requires the companion packages for test and convex optimization).