“…Thus if one wishes to construct an algorithm that computes some S and U of (0.1) at y 0 , a different method is needed. In [37] an analytic, rather than algebraic, approach to the eigenvalue problem is used to allow the computation of S and U without calling for eigenvectors. The analytic approach is based on function-theoretical treatment of resolvents (see, e.g., [84], [46], [47], and [70]).…”