A class of outer generalized inverses (Q2427893)

The notion of the Drazin generalized inverse in associative rings and semi-groups, introduced by the author of the present article [Am. Math. Mon. 65, 506--514 (1958; Zbl 0083.02901)] is as follows: Let \(S\) be a semigroup. An element \(b \in S\), is called the Drazin inverse of \(a\) if \(ab=ba,b=bab\) and \(a^{k+1}b=a^k\) for some nonnegative integer \(k\). Such a \(b\) is unique and is denoted by \(a^D\). Let us also recall the more widely known generalized inverse. This time, let \(S\) be a multiplicative semigroup with a given involution \(*\). For \(a \in S\) it is the Moore-Penrose inverse \(a^{\dagger}\) which is defined as the (unique) element \(b \in S\) that satisfies the equations: \(aba=a, bab=b; (ab)^*=ab\) and \((ba)^*=ba\). Any \(b \in S\) that satisfies the equation \(bab=b\) is called an outer inverse of \(a \in S\). Clearly, the two generalized inverses given earlier are specific outer inverses. In the paper under review, the author outlines a unified theory emcompassing a large class of uniquely-defined outer generalized inverses, especially including the Moore-Penrose inverse and the Drazin inverse. This theory is developed by exploiting the similarity and differences between these two classes of outer generalized inverses.