A conjecture on \(B\)-groups (Q1955719): Difference between revisions

Given a finite group \(G\), let \(B(G)\) be the Burnside algebra of \(G\) over the field of rational numbers. The latter is a split semisimple commutative algebra, whose primitive idempotents \(e_H^G\), where \(H\) runs through a set of representatives of the conjugacy classes of subgroups of \(G\), are described by \([X]=\sum_H |X^H|e_H^G\), for all \(G\)-sets \(X\). Given a normal subgroup \(N\leq G\), the functor \(X\rightarrow N\backslash X\) from \(G\)-sets to \(G/N\)-sets induces the deflation homomorphism \(\text{Def}_{G/N}^G: B(G)\rightarrow B(G/N)\). Then, if \(N\) is maximal such that \(\text{Def}_{G/N}^G(e_G^G)\neq 0\), according to earlier work of \textit{S. Bouc} [J. Algebra 183, No. 3, 664--736 (1996; Zbl 0858.19001)] the quotient \(\beta(G):=G/N\) is unique up to isomorphism, and \(G\) is called a \(B\)-group if \(\beta(G)=G\). Now, motivated by a recent result of \textit{M. Baumann} [J. Algebra 344, No. 1, 284--295 (2011; Zbl 1242.19001)], the purpose of the present note is to promote the conjecture that \(\beta(G)\) is nilpotent if and only if \(G\) is nilpotent. As evidence, it is proved that this indeed holds true if \(G\) is assumed to be solvable.