Dynamic Modeling of CIM Systems

Dynamic Modeling is not a precise deened concept. Nevertheless , its ambiguity makes it appropriate to use in discussions about requirements for a modern modeling methodology. \Real" systems (as CIM systems are) tend to be complex and dynamic. Models of such systems (or better: of parts of such systems) should serve diierent purposes: They should allow to answer certain questions about the underlying system and it should be possible to integrate/combine/interconnect the models to answer more sophisticated questions or questions on a diierent level of abstraction (e.g., strategic management questions if the underlying models are models for supporting control decision on an operational level). \Dynamic" modeling might emphasize mainly three aspects: (1) the utilization of the model for representing systems (or parts of systems) with certain dynamic properties, (2) the embedding of the method into an iterated (and dynamic) modeling process, and (3), as CIM systems are \living" systems 1 , the system of models must be able to consistently capture changes of the underlying system's organization and consequences of changed modeling purposes/goals. These are the components of the dynamics of modeling, as discussed here. The article presents Petri nets as one possibility to provide answers to inquiries into dynamic systems, to embed models into a coherent modeling process, to react on changes of perspective and organization, and to integrate models over diierent levels of abstraction to support operational, tactical and strategic decision making.