Automating the Assembly of Security Assurance Case Fragments

Trustworthy agent-based simulation: the case for domain-specific modelling languages

“…An interesting research question in the area of simulation engineering is, then, which aspects of the fitness-for-purpose argument can be instantiated automatically, and which require human construction. [ 29 ] also introduces the interesting idea of defeaters as part of the assurance case: these “capture doubts and objections” and may be useful in simulation engineering to capture, for example, calibration failures that motivated changes in the model and other information about the incremental and iterative refinement of the model.…”

“…However, some aspects will be largely mechanistic (e.g., connecting to appropriate data sources and establishing relevant properties), while others require more significant human input. In the safety assurance literature, this has been recognised early [14], leading to significant research into (often model-driven) mechanisms for automat-ically instantiating assurance cases (e.g., [18,19,29]). This is typically based on instantiating assurance patterns, explicitly leaving open inputs that need to be provided by humans; offering a way to focus human attention and drive system design based on where the most value can be added to the overall case [48].…”