Model Order Reduction: a comparison between Fractional and Integer Order Approximation

Abstract

The paper reports and compare some strategies to reduce high integer order transfer functions, exploiting both integer and non integer order calculus. In particular, the fractional order approach based on the implicit model, a genetic algorithm based procedure and the open loop balanced realization have been compared. Being fractional order system described in terms of pseudo-space, the number of the parameters of the reduced order models and their frequency response errors have been taken as performance indexes.