Transmission of continuity to the Bohr topology (Q1356958)

Here is a question addressed in the 1950's and 1960's by A. Beurling and H. Helson, by P. J. Cohen, and by W. Rudin: To what extent does the convolution algebra \(L^1(G)\) (with \(G\) an LCA group) characterize \(G\)? (Alternative version of this problem: Given LCA groups \(G\) and \(H\), find necessary and/or sufficient conditions on \(G\) and \(H\) that the algebras \(L^1(G)\) and \(L^1(H)\) are isomorphic.) Building upon and improving results of the above-cited authors, the present authors show these two principal results. Theorem 3.3: Given \(G\) LCA and \(D\) a discrete torsion-free Abelian group, the algebras \(L^1(G)\) and \(L^1(\mathbb{T}^m\times\mathbb{R}^n\times D\)) are isomorphic if and only if \(G\) is topologically isomorphic to \(\mathbb{T}^m\times \mathbb{R}^n\times D\times F\) for some finite Abelian group \(F\). Theorem 3.4: Torsion-free, compactly generated LCA groups \(G\) and \(H\) are topologically isomorphic if and only if \(L^1(G)\) and \(L^1(H)\) are isomorphic. The principal new insights making these results possible are (1) the concept of a piecewise affine map between LCA groups, (essentially) defined earlier by Cohen and by Rudin, and (2) the systematic exploitation in the present context of the Bohr topology of an LCA group. Specifically, the authors show that for each piecewise affine map \(\alpha: G\to H\) with \(G\) and \(H\) LCA there is a dense, open subset \(U\) of \(G^+\) (that is, of \(G\) in its Bohr topology) such that \(\alpha|U: U\to H^+\) is continuous.