High-performance algorithms and software for systems and control computations

For large-scale problems, it is extremely important to achieve the best computational efficiency, by exploiting any special structure, and by using the potential of modern high-performance computer architectures. Therefore, developing robust, and efficient algorithms for control systems analysis and design is an objective of primary importance. Improved versions of some basic algorithms for computer-aided control systems design have been recently developed. These algorithms have been incorporated in the latest version of the Fortran 77 subroutine library SLICOT. This paper comments on some of these algorithmic improvements. Several functional capabilities of SLICOT Library routines are briefly mentioned. Performance results for some components of the library, in comparison with equivalent computations performed by some MATLAB functions, included in the MATLAB nucleus or in the MATLAB control toolbox, are presented.