An introduction to operator polynomials (Q1187642)

This book is devoted to the study of polynomials, \[ L(\lambda) = A_0 + \lambda A_1 + \cdots + \lambda^ m A_ m, \] of a complex variable \(\lambda\) whose coefficients \(A_ 0,...,A_ m\) are bounded linear operators acting on a Banach space X. Such polynomials, which are referred to as operator polynomials, arise naturally in the theory of differential equations and boundary value problems, and in various branches of mathematical physics. The present book focusses the attention on two main themes, namely that of linearization and that of factorization. A bounded linear operator T acting on a Banach space Y is called a linearization of L(\(\lambda)\) over an open set \(\Omega\) if the spectrum of T lies in \(\Omega\) and \[ L(\lambda) \oplus I_{Z_ 1} = E(\lambda) [(\lambda -T)\oplus I_{Z_2}] F(\lambda),\quad \lambda\in\Omega, \tag{\(*\)} \] for some Banach spaces \(Z_ 1\), \(Z_ 2\) and some operator-valued functions E(\(\cdot)\) and F(\(\cdot)\) which are analytic on \(\Omega\) and such that E(\(\lambda)\) and F(\(\lambda)\) are invertible operators for each \(\lambda\in \Omega\). A linearization fixes the spectral data of the operator polynomial, but to solve the inverse problem uniquely further information is needed. The latter is provided by spectral pairs and spectral triples, which consist of a linearization and one or two additional operators, which may be constructed from the operator functions E(\(\cdot)\) and F(\(\cdot)\) appearing in (\(*\)). If L(\(\lambda)\) is monic, i.e., the leading coefficient \(A_ m\) is the identity operator on X, then a spectral triple (T,Q,R) on \({\mathbb{C}}\) is provided by the following three operator matrices: \[ T = \begin{pmatrix} \end{pmatrix}, \quad Q = \begin{pmatrix} \end{pmatrix}, R = \begin{pmatrix} I & 0 & 0 & \hdots & 0 \end{pmatrix} \] On the basis of the above approach, problems about operator polynomials may be analysed and solved in terms of their linearizations and the operators in the corresponding spectral pairs and triples. For example, questions concerning existence, perturbation and explicit representations of factorizations, common multiples and common divisors can be answered in geometric terms using invariant subspaces of the linearization. This spectral analysis of operator polynomials, which has been developed during the last 15 years with the author of the book among the main contributors, is the principal topic of the book. The first chapter presents various constructions of a linearization, proves its uniqueness up to similarity, and deals with other related notions of linearization. Chapters II-V contain the general spectral theory of monic operator polynomials (including that of selfadjoint ones) with applications to the study of common multiples and problems of stable factorization. Chapters VI and VII develop the general theory for non- monic operator polynomials. The last three chapters deal with common divisors and common multiples, resultant and Bezoutian operators (with applications to problems concerning separation of spectra) and Wiener- Hopf factorization. The development sketched in this book parallels to a certain extent the development in the theory of matrix polynomials initiated by Gohberg, Lancaster and Rodman at the end of the seventies. But there are many significant differences which are well explained in the present text. Much of the material presented here is based on work which was done in the last decade. About two third of the more than 170 references concern publications that appeared in the last 10-15 years. The book is well written and well organized. Parts of it could be used as text for a graduate course; each chapter has a set of exercises. Also a number of unsolved problems are discussed. The book also complements nicely the resent monograph of \textit{A. S. Markus} [Introduction to the spectral theory of polynomial operator pencils, Transl. Math. Monographs 71 (1988)], which deals mainly with the problem of completeness of eigenvectors and generalized eigenvectors for operators polynomials. Together the two books present a more or less complete picture of the main advances in the theory of operator polynomials made in the last decades.