Tile the group (Q670715)

At the beginning of the fifties \textit{P. Erdős} [``On a problem concerning congruence systems.'', Mat. Lapok, 4,122--128 (1952)] introduced the notion of the disjoint covering system of arithmetic progressions. The author investigates its group-theoretic generalization: Let \(\{H_i\}_{i=1}^s\) be subgroups of a given group \(G\), all of finite index. Suppose that we have the coset decomposition \(G=\cup_{i=1}^s g_iH_i\). The author says that this partition has multiplicity if there exist \(1\leq i_1\neq i_2\leq s\) such that \(H_{i_1}\) and \(H_{i_2}\) are of equal indices in \(G\). Motivated by the result of \textit{A. Soifer} [The mathematical coloring book. Mathematics of coloring and the colorful life of its creators. New York, NY: Springer (2009; Zbl 1221.05001), 7--8], \textit{M. Herzog} and \textit{J. Schönheim} [``Research problem No. 9'', Can. Math. Bull. 17, 150 (1974)] conjectured that a coset decomposition of a group of the above type has a multiplicity. The author (i) mentions that using a computer assisted verification based on decomposion of 1 as a sum of unit fraction all groups of orders smaller than 240 fulfill the Herzog-Schönheim conjecture (for an extension of this result cf. \textit{L. Margolis} and \textit{O. Schnabel} [Beitr. Algebra Geom. 60, No. 3, 399--418 (2019; Zbl 1458.11019)]), (ii) gives a sufficient condition for finite groups to satisfy the Herzog-Schönheim conjecture in terms of primes in the prime decomposition of the order of \(G\), and (iii) gives a variant of the proof by \textit{B. Novák} and \textit{Š. Znám} [Am. Math. Mon. 81, 42--45 (1974; Zbl 0279.10043)] of the original Mirsky, Davenport, Newmann and Rado one via group characters (unfortunately, he quotes Břetislav Novák as N. Břetislav).