A functional model for polynomially posinormal operators (Q5943201): Difference between revisions

A bounded linear operator \(T\) on a complex separable Hilbert space \({\mathcal H}\) is called totally \(P\)-posinormal for a polynomial \(P\) vanishing at zero, if for each \(\lambda\in \mathbb{C}\) there exists a constant \(M(\lambda)\) such that \(\|\overline P(T^*- \overline\lambda) h\|\leq M(\lambda)\|(T- \lambda)h\|\) for each \(h\in{\mathcal H}\) and if moreover the function \(\lambda\mapsto M(\lambda)\) is bounded on each compact subset of \(\mathbb{C}\); it is called pure if there exists no non-trivial reducing subspace on which it is normal. For such operators a \(L({\mathcal X})\)-valued distribution is presented as a complete unitary invariant, where \({\mathcal X}\) is a subspace of \({\mathcal H}\) connected with the self-commutator of \(T\) and the commutator of \((T- \lambda)^*\) and \(P(T- \lambda)\). Further a functional model is constructed generalizing the functional model for hyponormal operators of \textit{M. Martin} and \textit{M. Putinar} [J. Funct. Anal. 73, 297-323 (1987; Zbl 0635.47023)]. Conditions to imply totally \(P\)-posinormality are discussed for operator valued functions in terms of their power expansions at infinity.