Frequency weighted optimal Hankel-norm approximation of scalar linear systems

Abstract

This thesis addresses the problem of model reduction for scalar linear time-invariant systems via the use of the optimal Hankel-norm approximation problem. Frequency weighting is combined with optimality in the Hankel norm to obtain a frequency shaped approximation to a given linear system. This is accomplished by the solution of a modified optimal Hankel-norm approximation problem. Also presented is an error analysis for the frequency shaped approximation.