Finite direct sums controlled by finitely many permutations. (Q2393435)

Let \(R\) be a ring with \(1\neq 0\) and \(\mathcal S\) be a class of right \(R\)-modules. Let \(n\geq 1\). Finite direct sums of members of \(\mathcal S\) are said to be controlled by \(n\) permutations if we have the following: There are equivalence relations \(\mathcal R_1,\mathcal R_2,\dots,\mathcal R_n\), on \(\mathcal S\) such that given \(X_1,X_2,\dots,X_r\), \(Y_1,Y_2,\dots,Y_s\) in \(\mathcal S\), the direct sum \(X_1\oplus X_2\oplus\cdots\oplus X_r\) is isomorphic to \(Y_1\oplus Y_2\oplus\cdots\oplus Y_s\), if and only if \(r=s\) and there are \(n\) permutations \(\sigma_1,\dots,\sigma_n\) of \(\{1,2,\dots,r\}\) with \(X_\mu\mathcal R_iY_{\sigma_i(\mu)}\) for \(1\leq i\leq n\) and \(1\leq \mu\leq r\). In view of Krull-Schmidt-Remak-Azumaya theorem, [see \textit{A. Facchini}, Module theory. Endomorphism rings and direct sum decompositions in some classes of modules. Progr. Math. 167. Basel: Birkhäuser (1998; Zbl 0930.16001), Theorem 2.12], one immediately infers that (even infinite) direct sums of modules with local endomorphism ring are controlled by one permutation, where the only equivalence is the natural module isomorphism. Let us recall that a ring \(R\) is of type \(n\) if and only if \(R/J(R)\), where \(J(R)\) is the Jacobson radical of \(R\), is the direct product of \(n\) division rings and an object in a preadditive category is of type \(n\) if its endomorphism ring is of type \(n\) (e.g., a module whose endomorphism ring is local is of type \(1\) and it is assumed that the zero module is the unique module of type \(0\)), [see [12], \textit{A. Facchini} and \textit{P. Příhoda}, Commun. Algebra 39, No. 9, 3317-3338 (2011; Zbl 1237.16002)]. In [[13], \textit{A. Facchini} and \textit{P. Příhoda}, Algebr. Represent. Theory 14, No. 3, 545-570 (2011; Zbl 1259.16004], the question whether finite direct sums of members of a class \(\mathcal S\) of modules of type \(2\), are controlled by two permutations is fully studied. Motivated by the results and the techniques in [12, loc. cit., 13, loc. cit.], the author of the article under review investigates the latter question for a class \(\mathcal S\) of modules of type \(n\), and in Theorem 3.16 gives sufficient conditions for the finite direct sums of members of this class to be controlled by \(n\) permutations. Generally, the author works in a particular category of objects (still of finite type) and obtains a slightly more general result (namely, Theorem 3.14). Moreover, it is observed that [\textit{B. Amini, A. Amini, A. Facchini}, Commun. Algebra 39, No. 1, 76-99 (2011; Zbl 1225.16002), Theorem 5.3], is a consequence of this theorem. A slightly improved version of the classical Krull-Schmidt theorem (Theorem 2.2) which generalizes [13, loc. cit., Theorem 3.3] is also observed. Some examples of full subcategories of modules for which Theorem 3.14 applies are presented in Section 4. In the final section of the article, applying certain hypergraph associated to a class \(\mathcal S\) of modules, similar facts to Theorems 3.14, 3.16, are studied (Proposition 5.3, Theorem 5.4). It is observed that the latter theorem generalizes [13, loc. cit., Proposition 3.5].