Some differential identities on prime and semiprime rings and Banach algebras (Q2311405)

Let \(\mathcal{R}\) be a prime ring and let \(\mathcal{J}\) be a nonzero ideal. In [Result. Math. 42, No. 1--2, 3--8 (2002; Zbl 1038.16021)], \textit{M. Ashraf} and \textit{Nadeem-ur-Rehman} proved that \(\mathcal{R}\) is commutative if there exists a nonzero derivation \(d\) of \(\mathcal{R}\) satisfying \(d(g \circ h)= g \circ h\), for all \(g\), \(h \in \mathcal{J}\). In [\textit{N. Argaç} and \textit{H. G. Inceboz}, J. Korean Math. Soc. 46, No. 5, 997--1005 (2009; Zbl 1185.16044)], it was proved that it is enough to assume that \(d(g \circ h)^{n}= g \circ h\) holds true for some fixed positive number \(n\). In another direction, the original result of Ashraf and Rehman was extended to generalized derivations by \textit{S. Huang} [Taiwanese J. Math. 16, No. 2, 771--776 (2012; Zbl 1252.16038)]. In this paper, the authors show that \(\mathcal{R}\) is commutative when there exist positive integers \(m\) and \(n\), and a generalized derivation \(\mathcal{G}\), associated to an nonzero derivation \(d\), such that \(\mathcal{G}((g\circ h )^{2})^{m} = \mathcal{G}(g \circ h)^{2n}\). This and related results are also applied to describe some range inclusions in the case of continuous generalized derivations on Banach algebras.