On the \(G\)-compactifications of the rational numbers (Q2391063)

It is shown that for every sufficiently homogeneous metrizable compactum \(X\) there is a Polish group \(G\) acting on \(\mathbb Q\) such that \(X\) is homeomorphic to the unique \(G\)-compactification of \(\mathbb Q\). Here, `sufficiently homogeneous' can mean either that the homeomorphism group is transitive on points and the homeomorphism group of any point complement is transitive on countable dense subsets, or that every point has an open neighborhood \(B\) such that the homeomorphisms of \(X\) which act trivially on the complement of \(B\) are transitive on \(B\). If \(X\) is the Hilbert cube, it is shown that \(G\) is at most one-dimensional. This implies that there does not exist a dimension-preserving \(G\)-completion, anwering in the negative a question of Megrelishvili.