Common cyclic vectors for classes of operators having common spanning root spaces (Q860641)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Common cyclic vectors for classes of operators having common spanning root spaces |
scientific article |
Statements
Common cyclic vectors for classes of operators having common spanning root spaces (English)
0 references
9 January 2007
0 references
The research involving hypercyclicity of operators on Banach spaces has witnessed enormous interest in recent years, thanks not least due to the groundbreaking work by \textit{S.\,I.\thinspace Ansari} on the existence of a hypercyclic operator on each infinite-dimensional separable Banach space [J.~Funct.\ Anal.\ 148, No.\,2, 384--390 (1997; Zbl 0898.47019)]. There has also been considerable progress in the study of the closely related theory involving supercyclic and cyclic operators on Banach spaces. We recall that a bounded linear operator \(T:X\to X\) on a Banach space \(X\) is hypercyclic (resp., supercyclic, cyclic) if there exists \(x\) in \(X\) such that the set \(\{T^n (x): n\geq 1\}\) (resp., \(\{\lambda T^n (x): \lambda\in\mathbb R,\;n\geq1\}\), span\([(T^n x): n\geq1]\)) is dense in \(X\). In this case, the vector \(x\) is called a hypercyclic (resp., supercyclic, cyclic) vector of \(T\). One of the main results in the theory surrounding hypercyclicity pertains to the structure of the set of hypercyclic vectors of a hypercyclic operator. It turns out that this set is topologically big in the sense that it is always dense and that under certain mild extra conditions, it is even `algebraically big' in that it is possible to locate a closed infinite-dimensional subspace in this set of hypercyclic vectors. There are also recent results guaranteeing the existence of a common closed infinite-dimensional subspace corresponding to a given collection of hypercyclic operators on a Banach space. The problem pertaining to the existence of a dense set of supercyclic (resp., cyclic) vectors corresponding to a supercyclic (resp., cyclic) operator on a Banach space has attracted a lot of attention in recent years. As already mentioned, quite a good deal is already known about this aspect of the theory involving hypercyclic vectors. However, the analogous case of cyclic vectors remains largely unexplored. The purpose of the present paper is to show that various classes \(S\) of operators on a Banach space \(X\) having a common collection of `root spaces' whose union is total in \(X\) have a dense set consisting of cyclic vectors common to the members of \(S\). An important ingredient of the proof is the uniqueness theorem for series of the Wolff--Denjoy type which pops up naturally in the proof of the main theorem. The setting for the class of operators on a Banach space \(X\) with the stated property is provided by an arbitrary bounded sequence \(\{r_n\}\) of positive integers and a linearly independent collection \(\{e_n(i),\;1\leq i\leq r_n\}\), \(n\geq1\), of unit vectors from \(X\) which is linearly dense in \(X\). Letting \(X_n= \text{span}[e_n(i); \;1\leq i\leq r_n]\), \(n\geq1\), the desired class of operators \(T:X\to X\) verifying the conclusion of the theorem is chosen by demanding that there exist a bounded sequence \(\{\lambda_n\}\) of distinct complex numbers and a bounded collection \(\{a_n(i , j),\;n\geq1,\;1\leq i,j \leq r_n \}\) of complex numbers such that for all \(n\geq1\), \(T = \lambda_nI+ N_n\) on \(X\), where \(N_n\) is the operator on \(X_n\) which has a special representation w.r.t.\ the basis \(\{ e_n(i),\;1\leq i\leq r_n \}\), \(n\geq1\), in terms of the numbers \(a_n(i , j),\;n\geq1,\;1\leq i,j\leq r_n\). The special case of diagonal operators with the common denseness property studied previously by \textit{J.\,Wermer} [Proc.\ Am.\ Math.\ Soc.\ 3, 270--277 (1952; Zbl 0046.33704)] corresponds to \(X\) being a Hilbert space with \(\{e_n(i),\;1\leq i\leq r_n\}\), \(n\geq1\), as an orthonormal basis in \(X\). The paper concludes with some comments on generalizations of the main theorem obtained by relaxing some of these conditions involved in the statement of the main theorem. The case where the boundedness of the collection \(\{a_n(i , j),\;n\geq1,\;1\leq i,j \leq r_n \}\) is dispensed with is especially interesting.
0 references
common cyclic vectors
0 references
complete operators
0 references
block upper-triangular operators
0 references