Optimization and Systems Theory Seminar
Friday, May 28, 1999, 11.00-12.00


Professor György Michaletzky
Eötvös Lorand University,
Budapest, Hungary

Stochastic approach to Nevanlinna-Pick interpolation problem

In this talk we show that the recursive formulae in the Nevanlinna-Pick interpolation problem formulated for positive real - possibly matrix valued - functions can be obtained using a stochastic (geometric) argument, similar to the Levinson-Durbin algorithm. Based on this observation we characterise all possible solutions of the interpolation problem.

We also generalise the Kimura-Georghiou parametrization to the multivariate case and derive a short state-space form of this parametrization.

Calendar of seminars
Last update: May 26, 1999 by Anders Forsgren, anders.forsgren@math.kth.se.