Structured finite-dimensional controller design by convex optimization Original Research Article

We discuss how to solve rather general structured controller design problems by convex optimization. In our main contribution we develop, under a specific one-block hypothesis on just one subsystem in the feedback interconnection, a semi-definite programming algorithm that allows to design controllers whose McMillan degree is bounded in terms of the underlying plant description. As an additional goal, we reveal the close interplay between structural controller design and so-called multi-objective control problems in which one imposes structural performance specifications on the closed-loop system. Finally, we include various motivations for the relevance of extending our results in several directions.