The conceivable evolution of information technology will enable mechatronic systems with inherent partial intelligence. We call this kind of systems self-optimizing. Self-optimizing systems are able to react autonomously and flexible to changing environmental conditions. They are capable of learning and optimizing their behavior at run-time. This paper presents the paradigm of self-optimization and shows how to develop self-optimizing systems. Focus is the conceptual design phase. The result of this phase is the principle solution. In this paper at first we explain how to describe the principle solution in a domain-spanning way. This is of high importance because a holistic description of the principle solution constitutes the basis for the communication and cooperation between the engineers from different domains who are engaged in developing a self-optimizing system. At second we introduce solution patterns for developing principle solutions. Solution patterns enable the reuse of design knowledge so that the development process will be more effective. This paper focuses on solution patterns including self-optimizing solutions. How the methods work we explain by the complex magnetic linear drive of a shuttle.

This content is only available via PDF.
You do not currently have access to this content.