Extension and lifting of operators and polynomials (Q1762383)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Extension and lifting of operators and polynomials |
scientific article |
Statements
Extension and lifting of operators and polynomials (English)
0 references
23 November 2012
0 references
In what follows, \(\mathcal{A}\) is a class of operators, polynomials or holomorphic mappings between Banach spaces. Let \(X, Y, Z\) be Banach spaces, take a short exact sequence \(S:0\to Y\overset{i} \rightarrow X\overset{q} \rightarrow {Z} 0\) and say that \(S\) locally \(\mathcal{A}\)-splits [locally \(\mathcal{A}\)-lifts] if, for every Banach space \(E\), every \(t\in\mathcal{A}(E,Y)\) can be extended to a \(T\in\mathcal{A}(Y,E)\) through \(i\) (if, for every Banach space \(V\), every \(t\in\mathcal{A}(V,Z)\) can be lifted to a \(T\in\mathcal{A}(V,Z)\) through \(q\)). If the extension [lifting] \(t\to T\) can always be done with \(\|T\|\leq\lambda\|t\|\), then \(S\) uniformly \(\mathcal{A}\)-splits [uniformly \(\mathcal{A}\)-lifts]. The definition above unifies earlier definitions of splitting and lifting, and a main issue of the paper under review is indeed to unify the approach to the theory of splitting and lifting of exact sequences. Very interesting is the observation in Section 3 that BAP plays a key role concerning lifting: One may pass from \(\mathcal{F}\)-lifting to \(\mathcal{K}\)-lifting exactly when \(Z\) enjoys the BAP. Here, \(\mathcal{F}\) and \(\mathcal{K}\), as usual, denote finite rank and compact operators. Concerning splitting, one may always pass from \(\mathcal{F}\) to \(\mathcal{K}\). In Section 4, the situation for polynomial lifting and splitting is analysed; also here BAP is needed to pass from one class of polynomials to the next bigger one (Theorem 4). In Section 5, \(\mathcal{L}_\infty\)- and \(\mathcal{L}_1\)-spaces are characterized in terms of splitting and lifting, while in Section 6, the role of an operator ideal being balanced is remarked, and it is shown that the only classical injective and surjective operator ideal which is balanced is \(\mathcal{K}\). The paper is supplied with an appendix on formulations of many of the results in the language of homological algebra.
0 references
short exact sequence
0 references
lifting
0 references
splitting
0 references
exact functor
0 references
extension of operators
0 references