A new algorithm for RNO-INO type tensor product model representation

Abstract

A recently developed technique of modelling parameter varying dynamical systems is the Tensor Product model representation, where a system is decomposed to convex combination of several parameterinvariant models. Effectiveness and tractability of the representation depends strongly on the type of the applied convex combination. This paper presents an algorithm, which generates a special type of convex combination, the RNO-INO representation. An example is also introduced.