Exel's crossed product and crossed products by completely positive maps
Abstract: We introduce crossed products of a $C*$-algebra $A$ by a completely positive map $\varrho:A\to A$ relative to an ideal in $A$. They generalize various crossed products by endomorphisms when $\varrho$ is multiplicative. When $A$ is commutative they include $C*$-algebras associated to Markov operators by Ionescu, Muhly, Vega, and to topological relations by Brenken, but in general they are not modeled by topological quivers popularized by Muhly and Tomforde. We show that Exel's crossed product $A\rtimes_{\alpha,L} \mathbb{N}$, generalized to the case where $A$ is not necessarily unital, is the crossed product of $A$ by the transfer operator $L$ relative to the ideal generated by $\alpha(A)$. We give natural conditions under which $\alpha(A)$ is uniquely determined by $L$, and hence $A\rtimes_{\alpha,L} \mathbb{N}$ depends only on $L$. Moreover, the $C*$-algebra $\mathcal{O}(A,\alpha,L)$ associated to $(A,\alpha,L)$ by Exel and Royer always coincides with our unrelative crossed product by $L$. As another non-trivial application of our construction we extend a result of Brownlowe, Raeburn and Vittadello, by showing that the $C*$-algebra of an arbitrary infinite graph $E$ can be realized as a crossed product of the diagonal algebra $\mathcal{D}_E$ by a `Perron-Frobenious' operator $L$. The important difference is that in general there is no endomorphism $\alpha$ of $\mathcal{D}_E$ making $(\mathcal{D}_E,\alpha,L)$ an Exel system.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.