SVEP and generalized Weyl's theorem (Q2462216)

An operator \(T\) on a Banach space \(\mathcal X\) has the single-valued extension property at \(\lambda_0\in{\mathbb C}\) (SVEP for short) if, for every open disc \(D\) centered at \(\lambda_0\), the only analytic function \(f\colon D\rightarrow {\mathbb C}\) which satisfies \((T-\lambda)f(\lambda)=0\) for all \(\lambda\in D\) is the function \(f\equiv 0\). The quasinilpotent part \(H_0(T-\lambda)\) and the analytic core \(K(T-\lambda)\) are defined as \[ \begin{aligned} H_0(T-\lambda)&=\{x\in{\mathcal X}\mid \lim_{n\to\infty}\|(T-\lambda)^nx\|^{1/n}=0\}, \\ K(T-\lambda)&=\{x\in{\mathcal X}\mid\exists \{x_n\}\subset{\mathcal X} \;\text{and} \;\delta>0 \;\text{such that} \;\\ & x=x_0, \;(T-\lambda)x_{n+1}=x_n, \;\| x_n\|\leq \delta^n\| x\| \;(n=0,1,2,\ldots )\}. \end{aligned} \] An operator \(T\) satisfies a generalized Weyl condition (\(gWc\) for short) if \(\sigma(T)\setminus\sigma_{Bw}(T)= E(T)\), where \(E(T)\) denotes the set of isolated points of \(\sigma(T)\) which are eigenvalues and \(\sigma_{Bw}(T)\) is the set of complex numbers \(\lambda\) for which \(T-\lambda\) fails to be \(B\)-Weyl. An operator \(T\) is \(B\)-Weyl if there exists a natural number \(n\) such that the induced operator \(T_n: T^n({\mathcal X})\rightarrow T^n({\mathcal X})\) is Fredholm and has index \(0\). Let \(\gamma(T)\) denote the reduced minimum modulus of the operator \(T\). In the paper under review, the author gives necessary and sufficient conditions for an operator \(T\) to satisfy generalized Weyl or generalized Browder (defined in a similar way as \(gWc\)) conditions. One of the obtained results reads as follows: Theorem. If \(T\in B({\mathcal X})\) has SVEP at points \(\lambda\not\in\sigma_{Bw}(T)\), then \(T\) satisfies \(gWc\) if and only if one of the following equivalent conditions is satisfied: (a) There exists \(n\in{\mathbb N}\) such that \(H_0(T-\lambda)=(T-\lambda)^{-n}(0)\) for ever \(\lambda\in E(T)\). (b) There exists \(n\in{\mathbb N}\) such that \((T-\lambda)^n({\mathcal X})\) is closed for every \(\lambda\in E(T)\). (c) There exists \(n\in{\mathbb N}\) such that \(K(T-\lambda)=(T-\lambda)^n({\mathcal X})\) for every \(\lambda\in E(T)\). (d) There exists \(n\in{\mathbb N}\) such that \(\gamma(T_n)\) is discontinuous at every \(\lambda\in E(T)\). (e) \(\inf\{n\colon (T-\lambda)^n({\mathcal X})=(T-\lambda)^{n+1}({\mathcal X})\}<+\infty\) at each \(\lambda\in E(T)\). (f) \(E(T)\) is contained in the set of poles of the resolvent of \(T\).