An iterative test for \(H\)-matrix (Q1971838)

A matrix \(A\) is called generalized diagonally dominant, if there is a diagonal matrix \(D\) with positive entries on its diagonal such that \(AD\) is strictly diagonally dominant. An iteration for the computation of a suitable matrix \(D\) is presented. It also terminates after finitely many steps if such a matrix does not exist. Although the authors do not refer to this fact, the iteration is strongly connected to the computation of the positive eigenvector of matrices with nonnegative entries. The quotient criterion provides upper and lower bounds for the extremal eigenvalue which tell whether the iteration may be terminated.