Component Based Visualization of DIET Applications

Rolf-Hendrik van Lengen, Paul Marrow, Thies Bähr, Hans Hagen, Erwin Bonsma, Cefn Hoile

In: Georges-Pierre Bonneau , Thomas Ertl , Gregory M. Nielson (Hrsg.). Scientific Visualization: The Visual Extraction of Knowledge from Data. Seiten 367-384 Mathematics and Visualization ISBN 3-540-26066-8 ISBN 2006.


Controlling distributed information in a complex information infrastructure requires novel and innovative information processing and management techniques. The Decentralised Information Ecosystem Technologies (hereafter, DIET) approach provides a software platform based on a lightweight, robust, adaptable, and scalable multi-agent system. DIET uses implicit forms of communication found in natural ecosystems as an analogy for computer-based distributed information management systems. The platform can be used to tackle a variety of information management applications in distributed and open real-world scenarios. However, the development, debugging, and monitoring of distributed applications is a complex task. Adequate visualisation techniques and tools are required to assist the software developer. This paper describes the visualisation platform developed on top of the DIET framework. The main purpose of the visualisation platform is the inspection and manipulation of ecosystem inspired applications built upon DIET. The platform provides fundamental visualisation components that can be easily linked together to build up a visual network in order to formulate complex visualisation tasks. Possible tasks can range from monitoring an individual agent to visualising the overall system behaviour. Due to this component based approach the effort for monitoring and debugging agent applications is considerably reduced.

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence