Hypothesis-driven experiment design in computer simulation studies

In simulation studies, the goal specifies the objective or purpose of the study and thus drives the entire experimentation process. Relevant experiments and respective experiment hypotheses are derived from the study's goal and the model's observed behavior provides evidence whether these hypotheses hold. Current assistance systems do not integrate research hypotheses. Thus, the researcher has to make important design decisions which limits both replicability and reproducibility of the results. In this paper, the process of simulation studies is systematized based on a formally specified hypothesis. By this means, the research hypothesis becomes the key element of the study and guides the entire process. Hypothesis-driven simulation studies allow for the automated design, execution, and evaluation of experiments based on specific research questions. This facilitates documentation and execution as well as replication of simulation studies.