Downloads & Free Reading Options - Results

In this paper, we study the sample complexity of probabilistic methods for uncertain systems. In particular, we show the role of the binomial distribution for some problems involving analysis and design of robust controllers with finite families. We also address the particular case in which the design problem can be formulated as an uncertain convex optimization problem. The second main contribution of the paper is to study a general class of sequential algorithms which satisfy the required specifications using probabilistic validation methods and, at each iteration of the sequential algorithm, a candidate solution is probabilistically validated. The results of the paper provide the sample complexity which guarantees that the obtained solutions meet some pre-specified probabilistic specifications.