LMILab Translator

LMILab Translator

Introduction:

LMILab is a Matlab toolbox which has functions to solve semi-definite programming (SDP) problems. It exploits block structure of the linear matrix inequality (LMI) constraints. This structure is common in many control problems and for these problems, the algorithm used by LMILab is efficient. However, in problems with additional structure (e.g. sparse blocks), LMILab can perform quite poorly. Many freely available SDP solvers exist and most take advantage of sparsity in the LMI constraints. This code translates from LMILab form to other freely available solvers. This allows the user to specify the LMI constraints using the standard LMILab commands ('lmivar', 'lmiterm', etc.). The translator can then be used to solve the optimization problem with any of the freely available solvers. This provides a transparent alternative to LMILab.

Downloading and Usage:

Create a directory to store the translator.
Download the following file and store it in the directory: lmitrans.zip. This contains 9 m-files and one text file.
Unzip the file ('gunzip lmitrans.zip') and the file
The file, README, gives information on using the translator. Typing 'help minsdp' at the Matlab prompt gives more information.
Two comments: This translator was written and tested with Matlab 6. It was tested briefly in Matlab 5, so it should work with this version as well. Second, some of the solvers have function names which overlap with existing Matlab toolboxes and with each other. The function, 'solverpath', contained in the zip file above can be used to add and remove solvers from the Matlab path to prevent problems from overlapping functions.
For a list of updates and revisions: Revisions

SDP Software:

The following is a listing of some software for semi-definite programming problems. This table taken from http://members.home.net/nathan-brixius/sdplinks.html . The translator is currently able to convert to any of the Matlab-based solvers listed below.

NAME TYPE AUTHORS

CSDP C B. Borchers

SeDuMi Matlab,MEX Jos F. Sturm

SDPA or SDPA C++ K. Fujisawa, M. Kojima, K. Nakata

SDPHA Matlab F.A. Potra, R. Sheng, N. Brixius

SDPpack Matlab,MEX F. Alizadeh, J-P A. Haeberly, M. V. Nayakkankuppam, M. L. Overton, S. Schmieta

SDPT3 or SDPT3 Matlab,MEX K.C. Toh, M.J. Todd, R.H. Tutuncu

SP Matlab,MEX L. Vandenberghe, S. Boyd

Acknowledgement:

This work was supported under the DARPA Software Enabled Control program (contract: F33615-99-C-1497). For more information on DARPA and the Software Enabled Control (SEC) program:

DARPA: http://www.darpa.mil/
SEC: http://www.darpa.mil/ito/research/sec/index.html

Feedback:

Contact pseiler@vehicle.me.berkeley.edu if you have any comments and/or questions. Feedback and suggestions are also welcomed.

Disclaimer:

The author of this program accepts no responsibility for damages resulting from the use of this software and makes no warranty, either express or implied, including, but not limited to, any implied warranty of fitness for a particular purpose. This software is provided as it is, and you, its user, assume all risks when using it.

NAME	TYPE	AUTHORS
CSDP	C	B. Borchers
SeDuMi	Matlab,MEX	Jos F. Sturm
SDPA or SDPA	C++	K. Fujisawa, M. Kojima, K. Nakata
SDPHA	Matlab	F.A. Potra, R. Sheng, N. Brixius
SDPpack	Matlab,MEX	F. Alizadeh, J-P A. Haeberly, M. V. Nayakkankuppam, M. L. Overton, S. Schmieta
SDPT3 or SDPT3	Matlab,MEX	K.C. Toh, M.J. Todd, R.H. Tutuncu
SP	Matlab,MEX	L. Vandenberghe, S. Boyd