Consecutive-column and -row properties of matrices and the Loewner-Neville factorization (Q1372974): Difference between revisions

Consider matrices with entries in a ring with identity element. A matrix \(M\) is said to have consecutive-column property if for all \(k\) all its submatrices having \(k\) consecutive rows and the first \(k\) columns of \(M\) are invertible; consecutive-row property is defined similarly. Let \(B_i\) be a matrix with 1s along the diagonal, zeros outside the diagonal, and the subdiagonal and zeros also in rows \(2\) to \(n-i\) in the subdiagonal. Put \(B=B_1\cdots B_{n-1}\). The matrix C is defined similarly but with respect to the superdiagonal. Let \(D\) be a diagonal matrix. The factorization \(A=BDC\) is called Loewner-Neville factorization of the square matrix \(A\). As the main result in this paper the authors show, a square matrix \(A\) has both consecutive-column and consecutive-row properties if and only if it admits a Loewner-Neville factorization with all entries invertible, except those that were required above to be zero.