The realization problem of non-connected compacta as attractors (Q6077867)

Suppose that \(K\subseteq \mathbb{R}^n\) is a compactum which is a disjoint union of compacta \(K_1,\ldots , K_r\). The paper focuses on the following problem: {Problem.} \(K\) can be realized as an attractor for some dynamical system, either continuous (a flow) or discrete (a homeomorphism) if and only if each \(K_i\) can be so realized. The case of flows is solved thanks to the following result: {Theorem.} Let \(K\) be a disjoint union of compacta \(K_i\subseteq\mathbb{R}^n\). Then \(K\) can be realized as an attractor for a flow if and only if each \(K_i\) can be so realized. The case of discrete dynamical systems requires some extra effort. For this, the authors introduce the following notion. A finite family \(K_1,\ldots , K_r\subseteq\mathbb{R}^n\) of disjoint compacta is called ``totally splittable'' when there exist disjoint topological balls \(B_1, \ldots, B_r\) such that for each \(i\in\{1,\ldots ,r\}\), \(K_i\subseteq \textrm{int}(B_i)\). Then they obtain the next statement: {Theorem.} Let \(K\) be a disjoint union of a totally splittable family of compacta \(K_i\subseteq\mathbb{R}^n\). Then \(K\) can be realized as an attractor for a homeomorphism of \(\mathbb{R}^n\) if and only if each \(K_i\) can be so realized. An example is given to illustrate that in the absence of total splittability, the ``if'' part in previous theorem may fail.