Motivation. Simulation becomes more and more important in application areas, establishing itself as the third way of science in addition to theory and (real) experiments. To answer research questions, simulation studies form increasingly intricate processes that intertwine the design and execution of various, often calculation-intensive simulation experiments, the generation and refinement of simulation models, and steps of analysis.

The Dagstuhl Seminar has been dedicated to addressing central methodological challenges in supporting the conduction of effective and sustainable simulation studies. Thereby, the seminar focused on problems and solutions related to improving:

• Effectiveness: the usage of resources, including computing infrastructure and data, and the assistance of humans throughout a simulation study.
• Sustainability: continuing a simulation study into the future through support for reusing or building upon its central products, such as simulation model, data, and processes as well as the software used.

The last decades have seen a wide range of methodological developments in computer science that are likely to be instrumental in achieving effective and sustainable simulation studies. However, those efforts are scattered across different computer science fields that include high-performance computing, (modeling) language design, operations research, visual analytics, workflows, provenance, and machine learning, as well as modeling and simulation. The seminar brought participants with diverse computer science backgrounds together to enhance the methodological basis for conducting simulation studies.

Organization and results. Being one day shorter than typical seminars, the seminar started on Tuesday with a short round of introduction and continued with collecting ideas about achievements and challenges of modeling and simulation from the participants on 2 pinboards (see Figure 1). 3 talks and partly extensive discussions followed, one focusing on modeling and model-based approaches applied to simulation studies, one on high-performance computing for simulation, and one on analysis and experiment designs. In the late afternoon, the information gathered on the pinboards was revisited. In the end 3 working groups formed to work towards state-of-the-art and open-challenges papers on the following topics:

• Intelligent Modeling and Simulation Lifecycle
• Policy by simulation: seeing is believing for interactive model co-creation and effective intervention
• Context, composition, automation, and communication: towards sustainable simulation studies

Among the application fields as diverse as cell biological systems, traffic systems, or computer networks, one application dominated the discussions, i.e., Covid 19 simulation. The Covid pandemic showed the importance of modeling and simulation studies being conducted in an efficient, reliable manner, and, accordingly, of comprehensive, intelligent computer support for these studies, it revealed limitations, including those referring to communicating effectively modeling and simulation studies and their results to decision-makers. The results of the working groups are included as short summaries in this report. Wednesday afternoon, the participants presented their current research work and ideas in a series of lightning talks whose abstracts are also included in the report. However, most of the time was dedicated to the working groups. Plenary sessions on Thursday and finally on Friday allowed the participants to catch up with ideas and the progress made in the different working groups.

Creative Commons BY 4.0

Adelinde M. Uhrmacher, Wentong Cai, Christopher Carothers, and David M. Nicol