Simulation

Simulation is the by far the most used analysis techniques, both in academic and in industrial contexts. Although verification by simulation is not exhaustive, it often uncovers design problems in the early phases of the design and for very complex systems it is to date the only viable approach. Given the complexity of Systems of Systems, simulation is envisioned in DANSE as a critical analytical instrument for SoS that cannot be neglected.

Characteristics of the simulation of SoS include the following:
  1. Integration and coordination of a large number of possibly heterogeneous constituent systems;
  2. Presence of significant uncertainties, modeled as non-deterministic and stochastic behavior; 
  3. Different time scales within the same model.

The integration and coordination of heterogeneous constituent systems is achieved in the DANSE simulation framework through the FMI (Functional Mockup Interface) component integration standard. Proposals for extensions of the standard to support stochastic and non-deterministic modeling will be developed in DANSE. Different time scales are handled by decoupling SoS evolution from the detailed simulation.

The DANSE simulation framework will support three basic use cases:
  1. The joint simulation of the SoS with its constituent systems;
  2. The evaluation of the performance of different SoS alternatives;
  3. The support for simulation-based higher-level analyses.

The joint simulation of the SoS with its constituent systems takes advantage of the FMI component integration standard to support constituent systems defined within different modeling tools, such as Simulink, Modelica-based tools, Rhapsody, etc. FMU (Functional Mockup Units) are generated for each constituent system from the corresponding modeling tool. The SoS overall integration is specified in UPDM and understood by the DANSE simulation framework, which accordingly integrates the SoS constituent systems FMUs.


The DANSE simulation framework implements an FMI-compliant scheduler for the coordination of the constituent systems FMUs. The DANSE simulation framework supports the evaluation of the performance of different SoS alternatives. Different SoS architectural alternatives along with their corresponding executable models are specified or automatically generated in UPDM. User defined metrics for the evaluation of performance are also specified in UPDM, possibly reusing predefined metrics from the DANSE library of components. The DANSE simulation frameworks evaluates the metrics by executing the several joint simulations for the different SoS alternatives. The metrics are organized in two levels:
  1. Metrics evaluated over each single simulation;
  2. Metrics evaluated over multiple simulations of the same SoS.

The DANSE simulation framework defines a platform API for the support for simulation-based higher-level analyses. This API allows using the DANSE simulation framework for complex analyses independently from the specifics of the simulation framework implementation. This API is used in particular in DANSE to support the statistical verification of the SoS through statistical model checking.