Towards Liquid Models: An Evolutionary Modeling Approach

Today, we recognize a discrepancy between design time models concentrating on the desired behavior of a system and its real world correspondents reflecting deviations taking place at runtime. In order to close this gap, design time models must not be static, but evolutionary artifacts so called liquid models. Such liquid models are the cornerstone of our future research project CDL-MINT: Model Integrated Smart Production. In this position paper, we present an early result of this project: the liquid models architecture for linking design models to runtime concerns, which are derived from distributed and heterogeneous systems during operation. We elaborate on the proposed technologies for the respective architecture layers and identify the research challenges ahead.