More on the Bruhat order for (0, 1)-matrices (Q869891)

Let \(R = (r_1,r_2,\dots,r_m)\) and \(S = (s_1,s_2,\dots,s_n)\) be positive integral vectors. Then \(A(R, S)\) denotes the class of all \(m \times n\) \((0,1)\)-matrices with row sum vector \(R\) and column sum vector \(S\). Partial orders called Bruhat order and secondary Bruhat order were defined on \(A(R, S)\) in [\textit{R. A. Brualdi} and \textit{S.-G. Hwang}, ``A Bruhat order for the class of \((0,1)\)-matrices with row sum vector \(R\) and column sum vector \(S\)'', Electron. J. Linear Algebra 12, 6--16 (2004; Zbl 1088.05019)] and it was conjectured that these two are the same. In this paper, the authors construct a counterexample which shows that, in general, the Bruhat order is a proper refinement of the secondary Bruhat order. The counterexample consists of two matrices \(A, A'\) which have the property: \(A\) and \(A'\) are both minimal in the secondary Bruhat order but \(A\) strictly precedes \(A'\) in the Bruhat order. They proceed to show further that, in \(A(n,2)\), the Bruhat order is always the same as the secondary Bruhat order but, in \(A(n,3)\), they are not always so. (\(A(n,k)\) stands for the set of \(n \times n\) \((0,1)\)-matrices whose column and row sums are all equal to \(k\)). At the end, the authors mention six questions for further investigation.