Interpolation sets and the Bohr topology of locally compact groups (Q1882953)

Let \(G\) denote a locally compact (LC) group that is maximally almost periodic (MAP), that is, such that for every \(g \in G\) different from the identity \(e_G\) of \(G\), there is a continuous homomorphism \(f\) from \(G\) into a compact group \(K\) such that \(f(g) \neq e_K\). Then there exist a compact group \(bG\) and a continuous (necessarily 1-1) homomorphism \(r:G \rightarrow bG\) such that if \(f:G \rightarrow K\) is a continuous homomorphism into a compact group \(K\), then there exists \(\phi:bG \rightarrow K\) with \(f = \phi \circ r\). For historical reasons \((r,bG)\) is called the Bohr compactification of \(G\). Let \(G^+\) denote the underlying group of \(G\) equipped with its Bohr topology, i.e., \(G^+:=r[G] \subseteq bG\). Motivated by a result of \textit{I. Glicksberg} [Can. J. Math. 14, 269--276 (1962; Zbl 0109.02001)] in the sense that for \(G\) LC and Abelian, \(G\) and \(G^+\) have the same compact sets, there has been an extensive research studying the Bohr topology. But as far as the reviewer knows, the paper under review gives the first characterization of those LCMAP groups \(G\) with the property that \(G\) and \(G^+\) have the same compact sets. The present work draws from former contributions of the authors. Recall that a subset \(S\) of \(G\) is totally bounded if given any open subset \(U\) of \(G\), there exists a finite subset \(F\) of \(G\) such that \(S \subseteq FU\). The subset \(S\) of \(G\) is an interpolation subset if \(S\) is discrete as a subset of \(G^+\) and any bounded real-valued function of \(S\) can be extended continuously to \(bG\). Interpolation sets were introduced by \textit{S. Hartman} and \textit{C. Ryll-Nardzewski} [Colloq. Math. 12, 23-39 (1964; Zbl 0145.32101)]. Theorem. If \(G\) is an LCMAP, then \(G\) and \(G^+\) have the same compact sets if and only if every non totally bounded subset of \(G\) contains an interpolation subset.