Inverse \(M\)-matrix completion problem with zeros in the inverse completion (Q1614106)

A partially specified matrix \(A\) is said to be combinatorially symmetric if \(a_{ij}\) is specified if and only if \(a_{ji}\) is specified. The completion problem considered is to find the unspecified elements such that its completion is the inverse of an \(M\)-matrix. In the combinatorially symmetric case, this was solved by \textit{C. R. Johnson} and \textit{R. L. Smith} [Linear Algebra Appl. 241-243, 655-667 (1996; Zbl 0979.15009)], and in the unsymmetric case by \textit{L. Hogben} [ibid. 282, No. 1-3, 145-160 (1998; Zbl 0938.15009)]. In this paper, the completion problem is solved in the unsymmetric case with the extra condition that the inverse of the completed matrix has zeros at the positions that originally contained unspecified elements. It turns out to be the completion with the maximal determinant.