A new approach in the computation of poles and zeros in large networks

Algorithms for the computation of poles and zeros of rational network functions are presented, which use only the values of these functions and their derivatives at: discrete complex frequency points. The algorithms are obtained as an extension of root-finding methods without requiring as these the numerical values of the numerator and denominator polynomials separately. The derivatives of the network functions are determined from the coefficients of interpolation polynomials. The algorithms are very well sulted for the analysis of large networks. Examples are included.