Splitting dense columns in sparse linear systems (Q803727): Difference between revisions

Let \(A=[S,D]\) where S and D denote respectively the sparse and dense columns of a matrix A. The author gives an efficient and robust method for solving \(AA^ Tx=b.\) The proposed method avoids the rank-deficiency problem that is common to the present algorithms and also makes an effective use of sparse matrix techniques. Applications to interior-point methods for linear programming and non-symmetric matrices are pointed out.