Runtime adaptation of technical systems

This book introduces a novel system design approach and the corresponding framework to enable capabilities like self-configuration and self-improvement for parametrisable systems at runtime. As a result of these capabilities, systems equipped with the framework as additional control mechanism are characterised by aspects like adaptivity and robustness. Besides the general system design, the book investigates the possibility of applying machine learning techniques to real-world applications - two novel variants of Learning Classifier Systems and Fuzzy Classifier Systems are developed. These modified machine learning techniques are integrated into the framework. Thereby, they take over the responsibility of the self-improvement tasks of the system. Applications from various domains (e.g. vehicular traffic, data communication, and function approximation) serve as test bed for the evaluation.

Study of Computer Science (M.Sc.) at Leibniz Universität Hannover, Germany and University of Bristol, UK. Dr.-Ing. degree in Computer Science in 2011 at Leibniz Universität Hannover, System and Computer Architecture Group (Prof. Müller-Schloer). Currently Senior researcher at Universität Augsburg, Chair of Organic Computing.