Academic publication: "Verlässlichkeit selbstoptimierender Systeme"
In workgroup 1 several safety concepts are presented and discussed concerning their applicability to self-optimizing systems. Furthermore, impulses for co-operations between subprojects are realized in this workgroup. This communication platform also opens up for external experts; on the one hand, to put the developed methods and concepts up for discussion and on the other hand, to get new suggestions for possible methods and concepts and transfer them to the paradigm of self-optimization.
In the first funding period a collective understanding was established, which serves as a basis for the communication within the CRC 614. The term dependability comprises - according to Laprie - the attributes safety, reliability, availability and confidentiality. Further, relevant safety standards were checked considering the applicability to self-optimizing systems and the risk-oriented safety evidence was introduced. The three successive parts of this evidence are the risk analysis, the risk assessment and the counteraction planning. The workgroup “safety and stability” rely in the subsequent activities and results on this mutually created basis.
The focus of the second funding period lay on the risks and the potentials of self-optimization. Self-optimization holds risks due to its non-deterministic behaviour. Mainly in the design phase possible risks should be uncovered and minimized. The safety evidence was integrated into the dependability analysis process for the assessment of the risk. Further, issues regarding the reconfiguration by means of hybrid models with discrete and continuous states are analyzed in the design phase. However, in the operating phase, methods are applied which increase the dependability of the self-optimizing components and use the full potential of self-optimization. The link between design and operating phase is presented by the multi-level dependability concept. This dependability concept serves as condition monitoring of a self-optimizing component and is divided into four risk levels. In this funding period the concept was augmented in that way, that self-optimizing components and systems are capable of incorporating dependability with a defined weight within the system of objectives.
Methods and concepts developed in the subproject are described in detail in the book „Dependability of self-optimizing systems – using full potential and avoiding risks” (in german). This book summarizes the work of the first two funding periods (cover depicted in figure).
Overall, the second funding period dealt with concepts and methods developed in the respective subproject, which increase the dependability of self-optimizing components. In the third funding period the dependability in interconnected systems will be in the foreground. It should be validated, if the discussed methods and concepts are sufficient to guarantee the dependability of self-optimizing interconnected systems. Therefore, the experiences gained with the current methods and concepts are used within the workgroup and serve as a basis for augmentations and enhancements in the design of dependable self-optimizing.
