On the Dinitz conjecture and related conjectures (Q1901032)

The Dinitz conjecture states the following: let \(n\) be a natural number and \(S_{i, j}\) sets with \(n\) elements each \((i, j= 1,2,\dots, n)\). Then elements \(a_{i, j}\in S_{i, j}\) can be chosen such that \((a_{i, j})\) is a partial Latin square. The paper explores certain aspects of the Dinitz conjecture and its relationship with other conjecture involving Latin squares and ordered bases of vector spaces. One result is the proof of the Dinitz conjecture in a special case. As the Dinitz conjecture has been proven in the meantime (see \textit{Fred Galvin} [The list chromatic index of a bipartite multigraph, J. Comb. Theory, Ser. B 63, No. 1, 153-158 (1995; Zbl 0826.05026)]), this part of the paper has become obsolete. Another result is a proof of the special case \(n= 3\) of the following conjecture attributed to Rota: whenever \(n\) bases \(B_1, B_2,\dots, B_n\) of an \(n\)-dimensional complex vector space are given, then these can be ordered \(B_i= (b_{i1}, b_{i2},\dots, b_{in})\) such that for each \(j\) the \(b_{ij}\), \(i= 1,2,\dots, n\), form a basis. A similarity with the Dinitz conjecture cannot be denied. The paper closes with certain results on asymptotic enumeration of row-odd and row-even Latin rectangles.