Identification of Hybrid Systems via Mixed-Integer Programming
Author(s): A. Bemporad, J. Roll, L. Ljung |
Conference/Journal: vol. AUT00-29 |
Abstract: Many successful tools have been proposed by several researchers for safety analysis, stability analysis, and control/scheduling synthesis for hybrid systems. Such tools rely on a hybrid model of the process under investigation. Whenever the hybrid model is not available because some parameters are unknown, or too complex to be handled, getting a simple hybrid model from data is an identification problem, which has not yet received much attention in the hybrid systems community. In this paper we focus on the identification of piecewise affine (PWA) systems. PWA systems are a well studied class of hybrid systems, that was also shown to be equivalent to arbitrary interconnections of logic and linear dynamic components. We present in detail an algorithm for identifying piecewise affine Wiener models from input-output data. Two notable properties of the algorithm are that the globally optimal solution is found, and that the complexity is polynomial with respect to the number of data. Further Information |
Year: 2000 |
Type of Publication: (04)Technical Report | |
Supervisor: | |
% Autogenerated BibTeX entry @TechReport { BemRol:2000:IFA_936, author={A. Bemporad and J. Roll and L. Ljung}, title={{Identification of Hybrid Systems via Mixed-Integer Programming}}, institution={}, year={2000}, number={}, address={}, month=oct, url={https://old.control.ee.ethz.ch/index.cgi?page=publications;action=details;id=936} } |