Relaxed Vector Fitting Algorithm for Rational Approximation of Frequency Domain Responses

Experience with the vector fitting algorithm for rational approximation of frequency domain responses has in some situations revealed problems with slow convergence. A typical example is the fitting of noisy responses. The problem is linked to a high frequency asymptotic condition of the nontriviality constraint used in the vector fitting formulation. Relaxation of the non-triviality constraint is shown to be an effective solution to this problem. Application to selected examples demonstartes a significant improvement of performance in terms of convergence speed and the accuracy of the final result. Copyright IEEE