Robust stability and convexity: an introduction (Q1340990)

This book addresses questions related to the robustness analysis of the stability properties of a family of linear systems with constant coefficients. From the point of view of stability theory, constant coefficient linear systems can be identified with their characteristic polynomials. Thus a family of such systems can be identified with a subset \(B\) of the Euclidean space \(\mathbb{R}^n\) via the set of parameters occurring in a polynomial equation. To each point of \(B\) one associates the spectrum (i.e. the set of eigenvalues of the linear system it represents). Given a subset \(\Omega\) of the complex plane, the generalized robustness analysis involves finding conditions on a subset \(B_1\) of \(B\) such that, if the spectrum of every element in \(B_1\) lies in \(\Omega\), then the same is true for every element in \(B\). Convexity of various underlying sets (e.g. \(B\) and \(\Omega\)) play a role in answering such questions. In these problems, typically \(B\) is convex and \(B_1\) is a subset of its `vertices'. The book begins with the `four vertex theorem' of Kharitonov on the `Box' polynomial families and develops the notion of principal vertices. In the subsequent chapters convex analysis is employed to compute stability radii. The author also studies the case of multi-affine polynomial families. The zero exlusion criterion is discussed. Applications to systems with commensurate delays and to uncertain time-delay systems are also given in the later chapters. The book has an extensive bibliography on the subject.