On relative compactness in \(K(X,Y)\) (Q713437): Difference between revisions

The purpose of this note is to clarify the relation among some sufficient conditions for relative compactness of sets in \(K(X,Y)\), the space of compact operators between two Banach spaces \(X\) and \(Y\). In the spirit of the classical Arzelà-Ascoli theorem, a result by \textit{T. W. Palmer} [Proc. Am. Math. Soc. 20, No. 1, 101--106 (1969; Zbl 0165.47603)] asserts that a subset \(H\subset K(X,Y)\) is relatively compact if and only if \(\{T^*(y^*): T\in H,~ \|y^*\|\leq1\}\) is relatively compact in \(X^*\), and \(\{T(x):T\in H\}\) is relatively compact for every \(x\in X\). Sufficient conditions for compactness in \(K(X,Y)\) have been given by \textit{F. Mayoral} [Proc. Am. Math. Soc. 129, No. 1, 79--82 (2001; Zbl 0961.47010)] when the space \(X\) does not contain isomorphic copies of \(\ell_1\). Motivated by this result and Palmer's theorem, the author gives another sufficient condition for compactness in \(K(X,Y)\) when \(X^*\) satisfies the Gelfand-Phillips property (i.e., \(M\subset X^*\) is relatively compact when \(\sup_{x^*\in M}|x_n^{**}(x^*)|\rightarrow0\) for each weak* null sequence \((x_n^{**})\subset X^{**}\)).