Killing of supports on graded algebras. (Q2498891)

Let \(A=\bigoplus_{g\in G}A_g\) be an algebra graded by a group \(G\), and let \(\mathcal U\) be a subset of \(G\). Let \(A_{\mathcal U}=\bigoplus_{u\in{\mathcal U}}A_u\), and define a multiplication \(\cdot\) on \(A_{\mathcal U}\) by \(a\cdot b=ab\) for homogeneous elements \(a,b\) such that \(\deg(a)\deg(b)\in{\mathcal U}\), and \(a\cdot b=0\) otherwise, and then extending by linearity. One main theme of the paper is to discuss whether \(A_{\mathcal U}\) inherits a structure of a \(G\)-graded algebra with multiplication \(\cdot\), and if this is the case, what is the connection between the module theories for \(A\) and \(A_{\mathcal U}\). A second theme is the following. If \(G\) and \(G'\) are groups, and \(\varphi\colon G'\to G\) is a map, when is it true that for every \(G\)-graded algebra \(B=\bigoplus_{g\in G}B_g\) with \(\text{supp}(B)\subseteq\text{Im\,}\varphi\), taking \(\widetilde B_\sigma=B_{\varphi(\sigma)}\) for every \(\sigma\in G'\), gives rise to a \(G'\)-graded algebra \(\widetilde B\), whose underlying algebra is the same as for \(B\)? Again, in that case, what is the connection between the graded module theories of \(B\) and \(\widetilde B\)? Precise conditions are given for \(\mathcal U\) and \(\varphi\), and it is shown that the categories of modules are functorially well connected. If \(G=\mathbb{Z}\), the results are applied to \(n\)-Koszul algebras.