When the intrinsic algebraic entropy is not really intrinsic (Q890483): Difference between revisions

Peters introduced a notion of entropy for automorphisms of discrete abelian groups, showing that in the case of an automorphism of a discrete abelian group, the algebraic entropy coincides with the familiar Haar measure-theoretic or topological entropy of the dual automorphism of the compact dual group [\textit{J. Peters}, Adv. Math. 33, 1--13 (1979; Zbl 0421.28019)]. Roughly speaking the key insight is that the entropy in the usual dynamical sense on the compact group can be expressed in terms of a local volume decay of Bowen-Dinaburg balls at the identity, and the correct dual notion on the discrete group is an algebraic measure of expansion at infinity. One of the directions the study of these notions has taken was to introduce the intrinsic algebraic entropy for endomorphisms, using a counting argument similar to that used by Peters, but using the subgroups inert with respect to the endomorphism instead of the finite subgroups. Here it is shown that this intrinsic algebraic entropy can also be computed using the fully inert subgroups of sections of the group invariant under the endomorphism instead of the whole family of inert subgroups under the endomorphism. For a special class of groups it is shown that only fully inert subgroups of the group itself are needed to compute the intrinsic entropy, and examples are constructed to show that the same result does not hold in general.