Authors:
Jan Kantert
1
;
Sven Tomforde
2
;
Susanne Weber
1
and
Christian Müller-Schloer
1
Affiliations:
1
Leibniz University Hannover, Germany
;
2
University of Kassel, Germany
Keyword(s):
Multi-Agent-Systems, Testing, Self-organisation, Organic Computing.
Related
Ontology
Subjects/Areas/Topics:
Distributed Control Systems
;
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Intelligent Fault Detection and Identification
;
Software Agents for Intelligent Control Systems
Abstract:
Multi-agent systems typically consist of a large set of agents that act on behalf of different users. Due to
inherent dynamics in the interaction patterns of these agents, the system structure is typically self-organising
and appears at runtime. Testing self-organising systems is a severe challenge that has not received the necessary
attention within the last decade. Obviously, traditional testing methods reach their limitations and are hardly
applicable due to the runtime characteristics and dynamics of self-organisation. In this paper, we argue that
we run into a paradoxon if we try to utilise self-organising testing systems. In order to circumvent parts of the
underlying limitations, we propose to combine such an approach with instrumented testing.