From Domain-Spanning Conceptual Design to Domain- Specific Controller Design of Self-Optimizing Systems

Self-optimizing systems are technical systems with the ability to learn, communicate with each other, and optimize their behaviour autonomously in response to environmental changes. The operation of such systems requires close interaction between mechanics, electronics, control engineering and software engineering. This paper addresses the conceptual design of controllers based on the principle solution of self-optimizing systems. Controllers required by self-optimizing systems consist of multiple control algorithms. The control algorithms are designed to implement the control functionalities formulated during the conceptual design phase. The development of controllers based on the principle solution requires engineering skills not covered by control theory. These skills are extracting the control functionalities from the principle solution, decomposing the overall control problem and organizing the individual solutions to constitute a complete solution. A procedural model that integrates these skills is thus presented to link the domain-spanning principle solution phase with the doman-specific controller design phase.