Linear maps preserving idempotency of products or triple Jordan products of operators (Q927762): Difference between revisions

The authors discuss linear maps preserving idempotency of products or triple Jordan products of operators. Let \(\mathcal B(\mathcal X)\) be the algebra of all bounded linear operators on a complex Banach space \(\mathcal X\) and \(\Phi :\mathcal B(\mathcal X)\to\mathcal B(\mathcal X)\) a linear map. For \(A, B\in\mathcal B(\mathcal X)\), if \(AB\) is zero idempotent implies that \(\Phi(A)\Phi(B)\) is nonzero idempotent, then \(\Phi\) is called the nonzero idempotency of products of two operators; if \(ABA\) is zero idempotent implies that \(\Phi(A)\Phi(B)\Phi(A)\) is nonzero idempotent, then \(\Phi\) is called the nonzero idempotency of the triple Jordan product of two operators. It is proved that \(\Phi\) preserves the nonzero idempotency of products of two operators if and only if the following statements hold: {\parindent=6mm \begin{itemize}\item[(1)] If \(\mathcal X\) is finite-dimensional, then there exist an invertible operator \(A\in\mathcal B(\mathcal X)\) and a constant \(\lambda\in\{1, -1\}\) such that \item[(a)] \(\Phi(X) =\lambda AXA^{-1}\) for all \(X\in\mathcal B(\mathcal X)\); \item[(b)] \(n = 2\) and \(\Phi(X) = \lambda AX^t A^{-1}\) for all \(X\in\mathcal B(\mathcal X)\), where \(X^t\) denotes the transpose of \(X\). \item[(2)] If \(\mathcal X\) is infinite-dimensional and \(\Phi\) is surjective, there exists an invertible operator \(A\in\mathcal B(\mathcal X)\) and a constant \(\lambda\in\{1, -1\}\) such that \(\Phi(X) =\lambda AXA^{-1}\) for all \(X\in\mathcal B(\mathcal X)\). \end{itemize}} It is also proved that \(\Phi\) preserves the nonzero idempotency of the triple Jordan product of two operators if and only if the following statements hold: {\parindent=6mm \begin{itemize}\item[(3)] If \(\mathcal X\) is finite-dimensional, then there exist an invertible operator \(A\in\mathcal B(\mathcal X)\) and a constant \(\varepsilon\) with \(\varepsilon^3 = 1\) such that \item[(a)] \(\Phi(X) = \varepsilon AXA^{-1}\) for all \(X\in\mathcal B(\mathcal X)\); \item[(b)] \(\Phi(X) =\varepsilon AX^t A^{-1}\) for all \(X\in\mathcal B(\mathcal X)\), where \(X^t\) denotes the transpose of \(X\). \item[(4)] If \(\mathcal X\) is infinite-dimensional and \(\Phi\) is surjective, there exists a constant \(\varepsilon\) with \(\varepsilon^3 = 1\) such that either \item[(a)] there exists an invertible operator \(A\in\mathcal B(\mathcal X)\) such that \(\Phi(X) =\lambda AXA^{-1}\) for all \(X\in\mathcal B(\mathcal X)\); or \item[(b)] \(\mathcal X\) is reflexive and there is an invertible operator \(A\) from \(\mathcal X'\) onto \(\mathcal X\) such that \(\Phi(X) =\lambda AX A^{-1}\) for all \(X\in\mathcal B(\mathcal X)\), where \(\mathcal X'\) denotes the dual of \(\mathcal X\). \end{itemize}}