Authors:
Christian Becker
1
;
Jörg Hähner
2
and
Sven Tomforde
2
Affiliations:
1
Leibniz Universitaet Hannover, Germany
;
2
Augsburg University, Germany
Keyword(s):
Organic Computing, Flexibility, Machine Learning, Goal Exchange, Network Protocol Configuration.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Computational Intelligence
;
Evolutionary Computation and Control
;
Evolutionary Computing
;
Genetic Algorithms
;
Hybrid Learning Systems
;
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Machine Learning in Control Applications
;
Soft Computing
Abstract:
Within the last decade, technical systems that are capable of self-adaptation at runtime emerged as challenging approach to cope with the increasing complexity and interconnectedness in today’s development and management processes. One major aspect of these systems is their ability to learn appropriate responses for all kinds of possibly occurring situations. Learning requires a goal function given by the user – which is subject to modifications at runtime. In order to allow for flexible manipulations of goals within the system’s operation period, the learning component must be able to keep knowledge in order to respond to varying goals quickly. This paper describes attempts to implementing flexible learning in rule-based systems. First results show that efficient approaches are possible even in real-world applications.