The numerical range of a contraction with finite defect numbers (Q488987)

The paper is a study on the numerical range of a contraction operator with finite defect numbers. In the case when both defect numbers of a contraction \(T\) are equal, i.e., \(\dim{\mathcal D}_T=\dim{\mathcal D}_{T^*}=n\), it is proved that the closure of the numerical range of \(T\) is the intersection of the closures of the numerical ranges of its \(n\)-dilations, where an \(n\)-dilation of \(T\) on a Hilbert space \(\mathcal H\) is a unitary dilation acting on \({\mathcal H}\oplus \mathbb C^n\). The paper is organized in seven sections, the first one presenting some facts about numerical range and dilations. The second section deals with unitary \(n\)-dilations, and the main obtained result permits a detailed investigation of the properties of the numerical range. In the third section, the authors study the particular case when \(\dim{\mathcal D}_T=\dim{\mathcal D}_{T^*}=1\) for a completely nonunitary contraction \(T\), and the results give the possibility to obtain detailed information about the geometrical properties of the numerical range, extending previously obtained results for certain finite- or infinite-dimensional operators. In the fourth section, the main elements of the theory of the characteristic functions in the scalar case are recalled and necessary properties for the study of the numerical range are studied. The fifth section is devoted to the behaviour of the intersection with circular segments that corresponds to connected components of the boundary, in order to investigate the boundary of the numerical range. The functional model is used in the sixth section to describe boundary points when an inner function \(\theta\) is considered, and invariant subspaces for the model operator \(S_\theta\) are described. The last section of the paper is dedicated to some examples illustrating the obtained results.