Towards the Design of Self-Optimizing Mechatronic Systems: Consistency Between Domain-Spanning and Domain-Specific Models

Future mechanical engineering systems will consist of configurations of many highly distributed system elements with inherent partial intelligence. The complexity of self optimizing systems will grow enormously. The first major document of an engineering process is a set of models called the principle solution laying down the basic structure of the system. It is the foundation for the subsequent domainspecific refinement phases. One purpose of this set of specifications is to give the engineers of the involved domains a common understanding about the system. The subsequent domain-specific development refines this domainspanning model. We propose to use the principle solution also as a central point for keeping consistency between the refining models of the domain-specific development. Consequently, it is necessary to keep the principle solution consistent with the domain-specific refinement. Changes of the initial principle solution which result from iterative steps have to be propagated to the domain-specific models, as well as changes of the domain-specific models, which are relevant for the overall system design, have to be propagated. As an example, we illustrate the consistency between the principle solution and software specific models. Therefore, we look in more detail at the views environment, functions and active structure of the principle solution and at the structural and behavioural models of the software design. We also explain how this procedure can be automated by applying triple graph grammars to maintain consistency.