M. Eisenberg, D.Lehner, R. Sindelar, M. Wimmer: Towards Reactive Planning with Digital Twins and Model-Driven Optimization, in T. Margaria, B. Steffen, (eds): Leveraging Applications of Formal Methods, Verification and Validation, Practice, Lecture Notes in Computer Science, Volume 13704, Springer, Cham, 11th International Symposium, ISoLA 2022, Rhodes, Greece, October 22–30, 2022. Doi: 10.1007/978-3-031-19762-8_5

Digital Twins are emerging in several domains. They allow to connect various models with running systems based on bi-directional data exchange. Thus, design models can be extended with runtime views which also opens the door for many additional techniques such as identifying unexpected system changes during runtime. However, dedicated reactions to these unexpected changes, such as adapting an existing plan which has been computed in advance and may no longer be seen beneficial, are still often neglected in Digital Twins. To tackle this shortcoming, we propose so-called reactive planning that integrates Digital Twins with planning approaches to react to unforeseen changes during plan execution. In particular, we introduce an extended Digital Twin architecture which allows to integrate existing model-driven optimization frameworks. Based on this integration, we present different strategies how the replanning can be performed by utilizing the information and services available in Digital Twins. We evaluate our approach for a stack allocation case study. This evaluation yields promising results on how to effectively improve existing plans during runtime, but also allows to identify future lines of research in this area.

Towards Reactive Planning with Digital Twins and Model-Driven Optimization