What is a scenario and a scenario-based approach in the development and validation of complex systems?

What are the basic terms and concepts of scenario-based procedures in the development and validation of complex systems?

A scenario is the interactive play of dynamic and static objects and systems (also system-of-systems) in certain constellations, taking into account the relevant influencing factors, environmental conditions and system configurations. The objects and systems of the scenario are related in a temporally and spatially limited context.

Accordingly, the scenario description includes on the one hand the list of the objects, systems, constellations and system relationships involved. On the other hand, the values and ranges of the parameters and properties of the objects and systems as well as various initial and boundary conditions of the scenario are described. This is done by specifying distributions and ranges of the values of the respective parameters.

In a certain execution of a scenario with defined or given constellations, parameter values and configurations, we speak of a certain case, example or sample, regardless of whether it is

• an executed simulation (simulation case),
• a real test (test case) or
• the certain observation/measurement from the real world (real case).

The transition from the abstract, formal scenario description to the certain cases takes place via sampling methods for simulations and physical tests. Typical methods include Monte Carlo simulations or Design-of-Experiments.

If you take a snapshot of the object poses of a certain point in time within a specific case, you speak of a scene.

A situation consists a scene including the connections and motivations/intentions of the objects involved. A sequence of consecutive situations composes to a certain case, which may represent a specific scenario.

A use case abstractly describes a specific, functional aspect of a particular system in the user's external view of the system. Typically, the use cases are used to specify and formulate requirements of systems or to describe the (desired) capabilities of a system. Within a scenario, depending on the specific cases, certain use cases of a system are operational and in effect. Thus, scenarios can be used for the exemplary representation and validation of the abstract use cases of a system.

Within a scenario, there may be restrictions to a use case if you address certain assumptions or activity options regarding the system actions to it. Especially with "intelligent" systems, different action alternatives are tried out and evaluated in order to determine the best possible actions for a situation. A specific manifestation of this is the "incremental probabilistic simulation".

A given case can be evaluated and described in retrospect by suitable parameters, attributes and performance criteria (so-called key performance indicators, KPIs) a-posteriori. These assessments are only available for cases. If the evaluation criteria describe the system effects, then these are suitable as a basis for an effectiveness assessment. Hereby the system effects for a specific set of scenarios are evaluated quantitatively and comparatively.

In a scenario catalog or a scenario base different scenarios and cases are collocated in a uniform database, regardless of the source from which the data originates. The scenarios are described in a suitably parameterized and attributed manner, so that selected subsets of the catalog can be accessed according to the interested analyses, such as certain operational domains of a system for effectiveness evaluation.

In scenario mining one gains insights from the scenario base by the application of data mining methods.

In scenario-based development and validation, the desired/required, compliant behavior of a system is represented in a set of scenarios that are relevant for the (safe) use of the system. For this purpose, the surrounding and environmental conditions, the interactive behavioral patterns as well as the configurations and constellations of the objects and systems involved are usually varied in such a way that the desired and possible operation and operational conditions are sufficiently covered with relevant scenarios and at least span the "Operational Design Domain" of the examined system.

Scenario management includes the overall and comprehensive process of scenario-based development and validation from

• the design of experiments and definition of the relevant scenarios,
• the systematic generation and acquistion of the scenario data,
• the collocation and arrangement of the scenario data,
• the plausibility check and quality assurance of the data,
• the analysis and evaluation of the scenarios,
• the continuous improvement and adaptation of the systems up to
• the ongoing,  iterative repetition.