The complexity of future software systems, which are embedded in a dynamic environment, challenges traditional system design. With ever-increasing requirements and unforeseeable changes at runtime, complete a priori specification is hardly possible. Our focus is the systematic design of reliable autonomous systems incorporating descriptive modeling, testing, and formal verification to ensure their correctness despite autonomy. Further, we develop self-* algorithms for runtime adaption and optimization as well as techniques to incorporate uncertainties about a system’s environment in its decisions. Areas of application are flexible production plants, smart grid applications, quadrocopter swarms for environmental monitoring, among others.

?