An efficient and reliable algorithm for computing the singular subspace of a matrix, associated with its smallest singular values (Q1095586)

An algorithm for computing the singular subspace of a matrix corresponding to its smallest singular value is presented. The method is a modification of the singular value decomposition algorithm and can be three times faster than the classical one. There are three reasons for its high efficiency. First, the Householder transformations of the bidiagonalization are only applied on the base vectors of the desired singular subspace. Second, the bidiagonal is only partially diagonalized and third, a choice is made between QR and QL iteration steps. The new algorithm can be used in least squares applications and solving linear equations.