A several complex variables approach to sensitivity analysis and structured singular values
In various topics of control theory an important issue is that of optimization. For example, in sensitivity analysis of linear systems one is asked to find a system representation which minimizes some scalar function defined on the manifold of all realizations of a given transfer function. The known approaches to these problems are in most cases ad hoc and are often based on a case-by-case study, instead of applying general existence or uniqueness results for the solutions of the involved optimization tasks, valid for a wider range of problems. In this paper a systematic approach to the various optimization tasks in sensitivity analysis is developed, based on a few general results and notions from the analytic function theory in several complex variables. A recent result of Azad and Loeb on plurisubharmonic functions on complex homogeneous spaces enables us to solve the \(L^ p\)-sensitivity minimization problem. But the method is by no means restricted to this particular optimization problem. As a further illustration of our approach we show that the same techniques also lead to, e.g., new results on the convexity and critical point structure of upper bound functions for the structured singular value.
This page was built for publication: A several complex variables approach to sensitivity analysis and structured singular values
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4029274)
