A compact metric space that is universal for orbit spectra of homeomorphisms (Q409619)

Does there exist a compact metric space \(X\) of size \(\kappa\) which admits every possible homeomorphism allowed on the class of compact metric spaces of size \(\kappa\)? The above question (as stated by the authors) motivated writing of the paper under review. The question is answered for \(\kappa\) countably infinite (in the negative) and \(\kappa=\mathfrak{c}\) (in the affirmative, by an explicit construction). A related result for separable metric spaces of size \(\mathfrak{c}\) is also given, in the presence of some set theoretic assumptions. We explain their results in more detail in what follows. Let \(T:X\to X\) be a function on a set \(X\). Consider an equivalence relation on \(X\) by setting \(x\equiv y\) if and only if there are \(m,n\in\mathbb{N}\) such that \(T^m(x)=T^n(y)\). Call the equivalence classes of this relation the orbits of \(T\). Let \(O\) be an orbit of \(x_0\). \(O\) is an \(n\)-cycle if \(O=\{x_0,T(x_0),T^2(x_0),\dots,T^{n-1}(x_0)\}\) consists of \(n\) distinct points and \(T^n(x_0)=x_0\). \(O\) is said to be a \(\mathbb{Z}\)-orbit if there are distinct points \(\{x_j:j\in\mathbb{Z}\}\subseteq O\) such that \(T(x_j)=x_{j+1}\). If \(O\) is neither a \(\mathbb{Z}\)-orbit nor an \(n\)-cycle for some \(n\), then \(O\) is called an \(\mathbb{N}\)-orbit. An orbit spectrum of \(T\) is the sequence of cardinals \[ \sigma (T)=(\nu,\zeta,\sigma_1,\sigma_2,\sigma_3,\dots), \] denoting the number of \(\mathbb{N}\)-orbits, \(\mathbb{Z}\)-orbits, and \(n\)-cycles (for \(n=1,2,3,\dots\)) respectively. Note that if \(T\) is a bijection then \(\nu=0\). A space \(X\) with topological property \(\mathcal{P}\) is said to be universal for homeomorphisms of spaces with property \(\mathcal{P}\) if whenever a space \(Y\) with \(\mathcal{P}\) has a homeomorphism \(h:Y\to Y\) then there is a homeomorphism \(T: X\to X\) with \(\sigma(h)=\sigma(T)\). Section 1 contains a brief introduction. Section 2 contains preliminary definitions and results. In Section 3 (Theorem 3.1) the authors prove that no space is universal for homeomorphisms of countably infinite metric compacta. In Section 4 a universal compact metric space \(X\) for bijections that can be realized as homeomorphisms of metric compacta is constructed. \(X\) is a compact subset of a family of concentric circles in the complex plane \(\mathbb{C}\). Namely, let \(S_r=\{z\in\mathbb{C}:|z|=r\}\) and \(D=\{d_n:n\geq 1\}\) be a countable dense subset of \(S_1\). The space \(X\) is defined by \[ X=S_1\cup\bigcup_{n\geq 1}S_{1-1/2^n}\cup\bigcup_{n\geq 1}(\{(1+1/2^n+1/2^j)d_n:n+1<j\}\cup\{(1+1/2^n)d_n\}). \] In other words, \(X\) decomposes into the union of two spaces \(Y\) and \(Z\) such that \(Y\cap Z=S_1\), where \(Y\setminus S_1\) is a sequence of concentric circles converging to \(S_1\), and \(Z\setminus S_1\) is an isolated sequence of copies of the convergent sequence \(\{0,1,\frac{1}{2},\frac{1}{3},\frac{1}{4},\dots\}\) whose union is dense in \(Z\). In Section 5, assuming \(\mathfrak{c}<\aleph_\omega\), the authors conclude their paper with a construction of a universal space for homeomorphisms of separable metric spaces of cardinality \(\mathfrak{c}\).