SoS Modeling Approach

The DANSE modeling formalism focuses on the SoS level representation of the SoS. These view addresses concepts like the Capabilities, Goals, Constituent Systems and their relations rather than addressing system specific aspects like Ports or Parts. The System View is defined using the SysML/UML modeling elements and is adapted to the needs derived from the SoS View.
The SoS View takes Maiers five characteristics into account by supporting properties for geographical distribution, different roles depending on the operational scenario, the ownership relation and support of changes on different timescales.
The second major influence is the contract-based design paradigm which is applied not only on the system level but extended to the SoS level. Especially the interface between a Constituent System in the SoS view and its System representation in the System View illustrates the difference and commonalities of both views.

Constituent systems (CS) participating in an SoS take different Roles in the SoS. Even these roles can be differentiated into Infrastructure and Aggregate but more precise role descriptions are required to describe operational scenarios. These roles are characterized by the provided/required Capabilities of CS taking on such a role.
In addition to the capabilities assigned to the CSs, also contracts and goals define which system can take such a role in the SoS. Contracts are thereby similar used as in classical system engineering in terms of defining assumptions of the CS to the SoS and the physical environment and providing a promised behavior if the assumptions are fulfilled. This behavioral specification includes the capabilities but not the goals of a CS. To be more precise there are system level contracts as mentioned above and SoS level contracts which argue over the role of an individual CS in the SoS and the roles and relations of other systems in the SoS. At the SoS level the capabilities and goals are part of the role description of the participating systems. The SoS goals should represent the combination of the local goals of the CSs and conflicts of local goals should be solved by the structure of the SoS if they are not changeable. One example could be that the CS try minimizing their costs but also shall be able to perform the required actions. This conflict could be solved by ensuring that the requested actions are somehow fairly distributed to all CSs because in this setting the individual CS are able to benefit from cooperation with others.

The interface between the SoS and the CSs is therefore characterized by the ability of a system to participate the SoS and the role it takes in the SoS. The first could be simplified by saying that the system is compatible with the SoS and the second one by the role it should take is attractive for the system. The attractiveness is directly related to the goals of the system. If a system cannot achieve its goals in general it will not participate in the SoS. On the other hand on the SoS level there has to be a common understanding or a set of regulations which make the relations a CS has/should/could participate and the function it has to fulfill. This understanding may be only a local view to the SoS which could lead to problems in cooperating with the SoS. Therefore we assume that the local views of the CS in the SoS have to be taken into account and we will call it “world model” later on. This world model takes not only the physical environment of the CS into account but also the role and relations of other CS in the SoS. This view is relevant due to incomplete and anticipated data. That means that most information in this view is derived from communication, local observations or even trusted observations of others. At this point the concept belief comes into play which allows arguing about incomplete data in a way that misunderstandings can be explicitly modeled.

The SoS level contains the capabilities, (local and global) goals, relations and the world models of the CSs. The regulations on the SoS level are defined as SoS contracts which argue about the relations and roles of other CS as assumption and about the provided capabilities and actions of individual CS as promise. These assumption-promise-pairs are similar to the ones on the system level but these argue not (only) about interaction with the physical environment based on observable variables but about anticipated information about the other CSs, their relations, capabilities, local goals and interactions and the global goals of the SoS. The system level does not directly take the SoS into account except the observable data which is part of each communication for example.