Triple factorisations of the general linear group and their associated geometries. (Q486207): Difference between revisions

The authors investigate so called triple factorizations of groups. The interest in such factorizations goes back to papers on \(ABA\)-groups, for example results due to \textit{D. G. Higman} and \textit{J. E. McLaughlin} [Ill. J. Math. 5, 382-397 (1961; Zbl 0104.14702)]. They also play some role in coding theory. In this paper the authors deal with the group \(G=\mathrm{GL}(V)\), \(V\) a vector space of dimension \(d\geq 2\) over a field in characteristic \(p\). Hence let \(U\) be an \(m\)-dimensional subspace and \(W\) be a \(k\)-dimensional subspace, \(G_U\), \(G_W\) the stabilizer of \(U\) and \(W\), respectively. Let \(j=\dim(U\cap W)\). Then the authors prove that \(G=G_UG_WG_U\) if and only if \(\max(0,m+k-n)\leq j\leq\frac{k}{2}+\max(0,m-\frac{n}{2})\). This is related to similar factorizations of the Weyl group \(\Sigma_d\). In a further result the authors consider stabilizers of direct sums. Hence if \(G=\mathrm{GL}_{2k}(q)\), \(P_U\) is the stabilizer of an \(m\)-space \(U\) and \(P_Q\) the stabilizer of a decomposition \(V=V_1\oplus V_2\), where \(\dim(U\cap V_i)=k_i\), \(i=1,2\), \(k_1\leq k_2\). Then \(G=P_UP_Q P_U\) if and only if \(3k_2\leq k+1+m+k_1\) and \((q,m,k,k_1,k_2)\neq(2,1,1,0,0)\) (the latter is a real exception). These factorizations lead to nice coset geometries which are called collinearly complete by the authors.