Super-optimal Hankel-norm approximations

The computation of superoptimal Hankel-norm approximations is considered. The existence and uniqueness of the superoptimal solution are proved using simple state-space calculations. The approach is unlike the work of N.J. Young (1987), which is based on conceptual operator-theoretic constructions. In addition, the authors give a detailed analysis of pole-zero cancellations in the algorithm and a bound on the McMillan degree of the superoptimal solution, which generalize the results of D.J.N. Limebeer et al. (1989)