Matroid matching in pseudomodular lattices (Q1812751)

A matching in a set \({\mathcal H}\) of lines in a matroid \(M\) is a subset \({\mathcal G}\) of \({\mathcal H}\) such that \(r({\mathcal G})=2 | {\mathcal G} |\), where \(r\) denotes the rank function in \(M\). The matroid matching problem is to determine, for a given \({\mathcal H}\) and \(M\), the maximal matching in \({\mathcal H}\). A minimax formula for a maximal matching was found by \textit{L. Lovász} [Acta Sci. Math. 42, 121-131 (1980; Zbl 0449.51008)] when \(M\) is a projective space. Later on, the minimax formula was extended to the class of full algebraic matroids by \textit{A. W. M. Dress} and \textit{L. Lovász} [Combinatorica 7, 39-48 (1987; Zbl 0627.05016)]. The paper under review extends the minimax formula to the class of pseudomodular matroids defined by \textit{A. Björner} and \textit{L. Lovász} [Acta Sci. Math. 51, No. 3/4, 295-308 (1987; Zbl 0643.05023)].