Inner automorphisms and some their generalizations (Q2402165)

In this paper, an overview of some results for normal, quasi-inner, polynomial, and power automorphisms of groups is given. Each of these automorphisms generalizes the notion of an inner automorphism in certain direction. A couple of open problems is formulated. Some results on the structure of normal automorphism groups of braid groups and pure braid groups are proved. In particular, the author proves that each normal automorphism of \(B_n\), \(n\geq 3\), is inner. He uses the result by \textit{J. L. Dyer} and \textit{E. K. Grossman} [Am. J. Math. 103, 1151--1169 (1981; Zbl 0476.20026)]: \(B_n =gp(\mathrm{Inn}(B_n), \mu )\) where \(\mu\) inverts each standard generator of \(B_n\). Hence, to prove that each normal automorphism is inner the author just shows that \(\mu \) is not normal. Remark. The origin of the notion \textit{normal automorphism} comes from Galois theory. By definition, a normal automorphism maps each normal subgroup of finite index to itself. Many classical results use this definition. The author calls an automorphism \textit{normal} if it maps each normal subgroup to itself. Such an automorphism is called \textit{strongly normal}. This notion is much stronger. Usually, it allows to prove that a normal automorphism is inner much more easily than it can be done in the classical definition. Many results cited in the paper are much deeper than it follows from the author's statements.