The analysis of optimization based controllers (Q5935487)

A framework to analyze the stability and robustness of optimization-based controllers by using Kuhn-Tucker necessary conditions for optimality together with the known S-procedure of \textit{V. A. Yakubovich} [Sov. Math., Dokl. 11, 882-886 (1970); translation from Dokl. Akad. SSSR 193, 57-60 (1970; Zbl 0225.93023)] is developed. Analysis questions are formulated as implications between quadratic and linear equalities and inequalities that characterize the solution to the involved on-line optimization problem. A final application of the S-procedure provides sufficient conditions for nominal and robust stability in the form of widely used linear matrix inequalities. The results encompass the standard stability and robustness analysis for a variety of receding horizon control systems, including systems with polytopic and structured uncertainty. A simple numerical example illustrates the methodology.