Ideal invariant injections (Q323818)

Let \(\omega:=\{0,1,\dots\}\), \(\mathbb{Z}\) stands for the set of all integers, and id is the identity function on \(\omega\). By an ideal \(I\) on \(\omega\) we mean an ideal of subsets of \(\omega\) such that \(\omega\notin I\) and \(\{n\}\in I\) for all \(n\in\omega\). If \(I\) is an ideal on \(\omega\), then \(I^*\) denotes its dual filter \(\{\omega\setminus A:A\in I\}\).NEWLINENEWLINEIn this paper, the authors work with injections from \(\omega\) to \(\omega\). The set of all such injections is denoted by \(\mathbf{Inj}\). Fix an ideal \(I\) on \(\omega\) and let \(f\in \mathbf{Inj}\). They say that \(f\) is \(I\)-invariant if \(f[A]\in I\) for all \(A\in I\). And \(f^{-1}\) is \(I\)-invariant if \(f^{-1}[A]\in I\) for all \(A\in I\). If \(f\) and \(f^{-1}\) are \(I\)-invariant, then \(f\) is called bi-\(I\)-invariant.NEWLINENEWLINEThe authors organize the paper as follows. In Section 2, they focus on injections invariant with respect to countably generated ideals. In Section 3, they study injections invariant with respect to maximal ideals. In Section 4, they discuss injections invariant with respect to various ideals induced by submeasures on \(\omega\). They show that every increasing function is invariant with respect to ideals from a large class, however, it is not so for Erdős-Ulam ideals. In Section 5, they characterize increasing injections that are bi-invariant with respect to the classical density ideal \(I_d\) and the summable ideal \(I_{1/n}\). In Section 6, they show some applications of ideal invariant injections to ideal convergence of sequences.