A Framework for the Metamodeling of Multivariant Systems and Reactive Simulation Model Generation and Execution

Abstract

The simulation-based study of Cyber-Physical Systems or complex production systems leads often to a vast amount of system variants. Each system variant is characterized by a particular model structure and parameter settings, although system variants may also share common parts. There are two main approaches for modeling such a set of system variants. On the one hand, all variants are mapped in a big model with variation points and on the other hand variants are specified on a higher level of abstraction using a metamodel that is processed with appropriate transformation methods. This paper proposes an approach for modeling system variants using the System Entity Structure (SES) Ontology. It introduces new concepts and advances the SES by a procedural knowledge specification. Moreover it proposes a sofware infrastructure for an automated and reactive generation and execution of simulation models based on a SES in combination with a model base. Finally it refers to a prototype implementation within MATLAB/Simulink and forward-looking within Python.