BMIsolver is a MATLAB code for solving optimization problems with BMI (bilinear matrix inequality) constraints arising in feedback controller design (spectral abscissa, H2, H-infinity and mixed H2/Hinf optimization problems). It is implemented based on a method proposed in [1] which is a joint work with S. Gummusoy and W. Michiels. The code is developed by Tran Dinh Quoc at Department of Electrical Engineering (ESAT/SCD) and Optimization in Engineering Center (OPTEC), KU Leuven, Belgium (under the supervision of Prof. M. Diehl). The current version requires YALMIP as a modeling language. We recommend SEDUMI as an SDP solver.
Download BMIsolver
Third-party packages
- YALMIP - A MATLAB modeling language toolbox, download at http://users.isy.liu.se/johanl/yalmip/
- SEDUMI - An SDP solver, download at http://sedumi.ie.lehigh.edu/.
- COMPLeib - A collection of test problems. Download at http://www.compleib.de/
References
- Quoc, T.D., Gumussoy, S., Michiels, W. and Diehl, M.: Combining Convex-Concave Decompositions and Linearization Approaches for solving BMIs, with Application to Static Output Feedback (Tech. Report, July, 2011) [pdf]




