Passivity-preserving model reduction by analytic interpolation (Q998196)

The authors show that the passivity-preserving reduced-order model computed by the method of \textit{D. C. Sorensen} [Syst. Control Lett. 54, No. 4, 347--360 (2005; Zbl 1129.93340)] and of \textit{A. C. Antoulas} [ibid. 54, No. 4, 361--374 (2005; Zbl 1129.93304)] coincides (in the limit) with the central (maximum-entropy) Nevanlinna-Pick interpolant with interpolation points in the mirror-image of the selected spectral zeros. Then, by tuning the interpolation points and the spectral zeros, as prescribed by the theory, they obtain considerably better reduced-order methods. They also show that, in the multi-input/multi-output case, Sorensen's algorithm actually amounts to tangential Nevanlinna-Pick interpolation.