On the stability of \(\theta\)-derivations on \(JB^*\)-triples (Q2641614): Difference between revisions

Let \({\mathcal J}\) be a complex \(JB^*\)-triple, i.e., a complex Banach algebra with a continuous triple product \(\{\cdot,\cdot,\cdot\}:{\mathcal J}^3\to {\mathcal J}\) (satisfying several conditions). Let \(D,\theta:{\mathcal J}\to{\mathcal J}\) be \textbf{C}-linear mappings. Then, the mapping \(D\) is called a \(\theta\)-derivation on \({\mathcal J}\) whenever \[ D(\{xyz\})=\{D(x)\theta(y)\theta(z)\}+\{\theta(x)D(y)\theta(z)\}+\{\theta(x)\theta(y)D(z)\},\qquad x,y,z\in{\mathcal J}.\tag{1} \] The authors prove the stability of \(\theta\)-derivations. Namely, if \(f,h:{\mathcal J}\to{\mathcal J}\) are (in some sense) approximately \textbf{C}-linear and the condition (1) holds for them also approximately only, then \(f\) and \(h\) can be approximated by unique \textbf{C}-linear mappings \(D,\theta:{\mathcal J}\to {\mathcal J}\), respectively, such that \(D\) is a \(\theta\)-derivation on \({\mathcal J}\). Some general versions of such a stability result are presented as well as several applications. Replacing (1) by \[ D(\{xxx\})=\{D(x)\theta(x)\theta(x)\}+\{\theta(x)D(x)\theta(x)\}+\{\theta(x)\theta(x)D(x)\},\qquad x\in{\mathcal J}\tag{2} \] the so called Jordan \(\theta\)-derivation is defined. Clearly, each \(\theta\)-derivation is a Jordan \(\theta\)-derivation. The question whether the reverse implication holds true concludes the paper.