On accurate quotient singular value computation in floating-point arithmetic (Q2706293)

The authors propose a new algorithm for floating-point computation of the quotient singular value decomposition of an arbitrary matrix pair \((A,B)\in \mathbb{R}^{m\times n}\times \mathbb{R}^{p\times n}\). The algorithm proposed in the paper is an improvement and generalization of the tangent algorithm and it is designed to reduce a general pair \((A,B)\) to a (regular) pair with finite quotient singular values.