Algorithmic methods for bottom-up generation of system-level RF macromodels

During the design of electronic circuits and systems, particularly those for RF communications, the need to abstract a subsystem from a greater level of detail to one at a lower level of detail arises frequently. One important application is to generate simple, yet accurate, system-level macromodels that capture circuit-level non-idealities such as distortion. In recent years, computational ("algorithmic") techniques have been developed that are capable of automating this abstraction process for broad classes of differential-equation-based systems (including non-linear ones). In this paper, we review the main ideas and techniques behind such algorithmic macromodelling methods.