Linear conditions for positive determinants (Q1124879)

Linear conditions on rows of an \(n \times n\) matrix ensuring the positivity of its determinant are studied. It is shown that any weakest set of such conditions must involve at most \(2^{n-1}\) linear inequalities per row for the case when the matrix is strictly diagonal dominant by rows and has positive diagonal entries. Then it is found the new weakest set of conditions with only \(n^{2}\) linear inequalities which requires that for each row of a matrix the row mean must be positive and larger than all the off-diagonal entries in that row. The case of a \(3 \times 3\) matrix is discussed in more detail and a graphical interpretation of derived formulae is given.