A tool-box approach to computer-aided generation of reduced-order models

Technical systems can be characterized very often as beiing very complex and heterogeneous. Many simulators exist for the analysis of continuous, discrete, and hybrid systems. However, the first modeling steps in physical domains are not yet sufficiently assisted by CAD tools. Multi-ports and generalized KIRCHHOFFian networks proved to the powerful concept in a physically oriented modeling methodology. If the physical system is described by partial differential equations, discretization algorithms are applied to construct models described by ordinary differential equations or equivalent lumped network models. Their parametrization can be supported by various algorithms for numerical order reduction in frequency and time domain, symbolic and semi-symbolic analysis, approximation, and optimization which are comprised in a tool box. Modeling languages as VHDL-AMS, Verilog-AMS, SystemC-AMS, and Modelica will support the practical application of these approaches