“…The approximation problem (15) can be solved using the famous AAK-lemma, see [17]. In particular, it is well-known that γ 1 (r)=σ r+1 (G) where σ i (G), i =1 ,...,n, are the Hankel singular values of the linear operator G, and methods for computing a state-space realization (A r ,K r ,C r ) of the optimal G * r are available, see [18], [19].…”