Generalized Goldie dimension for an \(S\)-act. (Q1775335)

The paper contains several propositions and one theorem as follows: Let \(M\) be an \(S\)-act and \(I\) an index set. Then \(\text{gd}(M)=|I|\) if and only if \(M\) has an essential subact which is a 0-direct union of uniform subacts of \(M\) with \(|I|\) terms. Here the following definitions are used: `generalized Goldie dimension' \(\text{gd}(M)=|I|\) means that the zero disjoint union \(\bigcup_{i\in I}M_i\) of all maximal uniform subacts \(M_i\) of \(M\) is essential in \(M\). A subact \(N\) of \(M\) is called `essential' (or `large') in \(M\) written \(N\triangleleft M\), if \(N\cap K\not=0\) for each subact \(0\not=K\leq M\). A non-zero \(S\)-act \(N\) is called `uniform' if every non-zero subact of \(N\) is essential in \(N\). The following examples are given: Let \(T\) be a semigroup which composes of infinite orthogonal idempotents \(e_i\), \(i\in\mathbb{Z}^+\), where \(\mathbb{Z}^+\) is the set of positive integers. Let \(S=T^1\) be the semigroup obtained from \(T\) by adjoining the identity 1. Then \(G(_SS)=\coprod_{i\in\mathbb{Z}^+}Se_i\) and \(G(_SS)\) is the unique essential subact. Let \(I=2\mathbb{Z}^+\) and \(K=\coprod_{i\in I}Se_i\). Then \(K=G(K)\triangleleft K\) and hence \(\text{gd}(K)=|I|=|\mathbb{Z}^+|=\text{gd}(S)\). It is easy to see that \(K\) is not essential in \(S\). Another example shows that the converse of the statement ``If \(K\) is closed in \(M\), then \(Q/K\triangleleft M/K\) whenever \(Q\triangleleft M\) such that \(K\subset Q\)'', which has been established for subacts \(K\subset L,N\) of \(M\), is not true. Let \(S=\{0,1,2,\dots\}\) with multiplication on \(S\) as follows: for any \(m,n\in S\), we define \[ 0m=m0=0,\quad 1m=m1=m; \] and if \(m,n\not=0,1\), then \[ mn=nm=\begin{cases}\max\{m,n\}&\text{if }m-n\text{ is even},\\ 0&\text{otherwise.}\end{cases} \] Now, \(S\) together with the above multiplication becomes a commutative semigroup. Let \(K=S2\coprod S3\). Then \(K\triangleleft_SS\). Noting that \(_SS\) is the unique subact of \(_SS\) which properly contains \(K\), the conclusion of the statement holds. However, \(K\) is clearly not closed in \(_SS\). Here a subact \(K\) of \(M\) is called `closed in' \(M\) if, whenever \(L\) is a subact of \(M\) such that \(K\) is essential in \(L\), then \(K=L\). One example shows that the equality in \(\text{gd}(M)\leq\text{gd}(N)+\text{gd}(M/N)\) which has been established under certain conditions does not hold in general. Let \(\mathbb{Z}\) be the multiplicative semigroup of integers and \(\mathbb{Q}\) the set of rational numbers. Then \(M={_\mathbb{Z}\mathbb{Q}}\) is a \(\mathbb{Z}\)-act under the usual multiplication, with \(\text{gd}(M)=1\) and \(\text{gd}(M/\mathbb{Z})=\infty\). Remarks: Terminology is not always explained, \(\triangle_S\) probably means \(\triangle\) and \(\mathbb{Q}/K\) probably denotes the Rees factor act.