The NEOS Server offers PENBMI for the solution of linear and bilinear matrix inequality problems.
More information on PENBMI is available in this
The code can be obtained from
The algorithm used combines ideas of the exterior (penalty) and interior
(barrier) methods with the Augmented Lagrangian method.
It can be called from a C/C++ program, a Fortran program, or from Matlab as
in this implementation.
PENBMI was developed by
Michal Kocvara and
This solver was implemented by
Hans Mittelmann and executes at
Using the NEOS Server for PENBMI
The user must submit a BMI problem as a Matlab structure, as an ASCII
file as in the Sample Submission problems, as a locally generated structure
"pen", or as YALMIP m-files. See, the user's guide.
Here are two examples of structures in binary
is an m-file to generate random BMI instances.