Risk-Averse Control Problems in Model-Following Systems
- 915 Downloads
This chapter considers performance information in a linear-quadratic class of model-following control systems. The innovative idea is the fact that performance information with higher-order performance-measure statistics can improve control decisions for closed-loop system performance reliability but the controller design can also be computationally involved. Many of the results entail measures of the amount, value, and cost of performance information, and the design of model-following control strategy with risk aversion. It becomes clear that the topic of performance information in control is of central importance for future research and development of correct-by-design of high performance and reliable systems.
KeywordsRisk Aversion Feedback Gain Performance Information Cost Cumulants Admissible Initial Condition
- 1.Pham, K.D.: On statistical control of stochastic servo-systems: performance-measure statistics and state-feedback paradigm. In: Proceedings of the 17th International Federation of Automatic Control World Congress, pp. 7994–8000 (2008)Google Scholar
- 5.Hassibi, B., Kailath, T.: Tracking with an H-infinity criterion. In: Proceedings of the Conference on Decision and Control, pp. 3594–3599 (1997)Google Scholar
- 6.Yuksel, S., Hindi, H., Crawford, L.: Optimal tracking with feedback-feedforward control separation over a network. In: Proceedings of the American Control Conference, pp. 3500–3506 (2006)Google Scholar
- 7.Pham, K.D.: Statistical control paradigms for structural vibration suppression. Ph.D. Dissertation. Department of Electrical Engineering. University of Notre Dame, Indiana, U.S.A. Available via http://etd.nd.edu/ETD-db/theses/available/etd-04152004-121926/unrestricted/PhamKD052004.pdf. Cited 15 March 2012 (2004)