OPTIMIZATION-BASED DESIGN OF CONTROL SYSTEMS WITH UNCERTAIN PLANT

Abstract This paper presents an optimization-based design methodology for the worst case design of single-input, single-output control systems when the plant model contains both parametric and unstructured uncertainty . It is shown that computational complexity can be considerably reduced, at the expense of some tightening of the specifications, by replacing the original performance-specifying semi-infinite inequalities by simpler majorizations . The decomposition and majorization results are also used to determine whether a controller structure can satisfy a given design requirement.