Commutativity preserving maps (Q935377)

This is a well-written paper devoted to the problem of characterizing maps preserving commutativity, one of the most studied linear preserver problems. Let \({\mathcal A}\) and \({\mathcal B}\) be two (associative) algebras over a field \({\mathbb F}\). A map \(\Phi: {\mathcal A} \rightarrow {\mathcal B}\) \textit{preserves commutativity} if \(xy=yx\) implies \(\Phi(x)\Phi(y)=\Phi(y)\Phi(x)\). The map \(\Phi\) \textit{preserves commutativity in both directions} if the converse implication is also true. The author begins by explaining that the problem of commutativity preserving maps which are linear is connected with the study of Lie homomorphisms, while the case of arbitrary maps is motivated by problems in mathematical foundations in quantum mechanics. Commutativity preserving maps on spaces of matrices or operators have been considered by several authors. The author gives here a brief survey and some historical aspects of this research area, initiated by \textit{W. Watkins} [ibid. 14, 29--35 (1976; Zbl 0329.15005)] in the case of the matrix algebra \( M_n({\mathbb F})\), and pursuing by \textit{M. Omladič} [J. Funct. Anal. 66, 105--122 (1986; Zbl 0587.47051)] for the algebra \({\mathcal B}(X)\) of bounded linear operators on a Banach space \(X\). Applying the technique of functional identities, \textit{M. Brěsar} [Contemp. Math. 259, 93--109 (2000; Zbl 0967.16011)] succeeded subsequently to characterize bijective linear maps preserving commutativity for prime algebras. Then the author discusses the literature on commutativity preserving linear maps on matrix spaces in the absence of the surjectivity hypothesis and recalls also some results in the non-linear preservers of commutativity. Finally the author proves the following result describing the general form of injective continuous commutativity preserving non-linear maps on matrix algebras: Theorem. Let \(n\geq 4\) and \(\Phi: M_n({\mathbb C}) \rightarrow M_n({\mathbb C})\) be an injective commutativity preserving continuous map. Then there exist an invertible matrix \(T\in M_n({\mathbb C})\) and a locally polynomial map \(A \mapsto p_A(A)\) such that one of the following conditions holds:{\parindent=8mm \begin{itemize}\item[(i)] \(\Phi(A)=Tp_A(A)T^{-1} \;\; \text{ for all } A\in M_n({\mathbb C})\); \item[(ii)] \(\Phi(A)=Tp_A(A^t)T^{-1} \;\; \text{ for all } A\in M_n({\mathbb C})\); \item[(iii)] \(\Phi(A)=Tp_A(\overline{A})T^{-1} \;\; \text{ for all } A\in M_n({\mathbb C})\); \item[(iv)] \(\Phi(A)=Tp_A(A^*)T^{-1} \;\; \text{ for all } A\in M_n({\mathbb C})\), \end{itemize}} where \(A^t, \, A^*\) and \(\overline{A}\) denote the transpose of \(A\), the conjugate transpose of \(A\) and the matrix with complex conjugated entries \(\overline{A} = [\overline{a_{ij}}]\), respectively.