A scalable moment matching-based model reduction technique of linear networks

Abstract

In this paper we address the problem of model order reduction of linear network systems. Using Sylvester equation-based moment matching techniques, we propose a framework to compute families of parametrized reduced order models that achieve moment matching and preserve the structure of the to-be-reduced model of the network. Further, using balanced truncation techniques we also reduce the number of subsystems in the network. The result is a low-order approximation of the linear network system with a reduced number of subsystems that exhibit properties similar to the given network. This approach leads to a scalable modeling algorithm for large-scale networks, using specific features of the system, such as the dynamical interactions between subsystems and the concepts from the model order reduction field.