Matrix rings satisfying column sum conditions versus structural matrix rings (Q2564895): Difference between revisions

Let \(R\) be an arbitrary associative ring with an identity element, and let \(B=[b_{ij}]\) be a reflexive and transitive \(n\times n\) matrix with entries in \(R\). By reflexivity, \(B\) satisfies \(b_{ii}=1\) for all \(i\), and, by transitivity, \(b_{ij}=1\) and \(b_{jk}=1\) imply \(b_{ik}=1\). Thus \(B\) is an \(n\times n\) quasiorder and determines the quasiorder relation \(c_B\) on the set \(\{1,\dots,n\}\) defined by \(ic_Bj\) if and only if \(b_{ij}=1\). The subset \(M_n(B,R)\) of \(M_n(R)\) of all matrices with \((i,j)\)th entry equal to 0 if \(b_{ij}=0\) forms a ring which is called a structural matrix ring. On the other hand, there is an internal characterization of structural matrix rings in terms of a set of matrix units associated with a quasiorder relation. The author uses this internal characterization to derive isomorphisms between certain subrings of a complete matrix ring. He applies this result to matrix rings satisfying column sum conditions.