One of the main challenges in multi-agent systems is to develop tools and techniques for solving cooperative tasks in a dynamically changing environment.
The aim of the annual agent contest is to stimulate research in the area of multi-agent systems, to identify key problems and to collect suitable benchmarks.
MSc students Niklas Skamriis Boss & Andreas Schmidt Jensen (right) in front of a small sample scenario with cows (brown dots), cowboys (red and blue dots),
obstacles (green dots) and corrals with fences (red and blue areas).
The 2009 contest was organized by Tristan Behrens, Mehdi Dastani, Jürgen Dix, Michael Köster & Peter Novák - further information about the sample scenario and the contest is available here: http://www.multiagentcontest.org
The DTU team used the Jason platform and did well in the 21 world-wide simulations during 5 days:
1. JIAC-V (Germany) 60 points 2. Jadex@HAW (Germany) 57 points 3. Roman-Farmers (Brazil/France/Italy) 37 points 4. Jason-DTU (Denmark) 30 points 5. Smaperteam (Australia) 23 points 6. Micro-JIAC (Germany) 21 points 7. AF-ABLE (Ireland) 20 points 8. Unknown (Germany) 1 point
Jason-DTU source code: ZIP
Niklas Skamriis Boss, Andreas Schmidt Jensen & Jørgen Villadsen:
Developing Artificial Herders Using Jason
Pages 193-197 in Jürgen Dix, Michael Fisher & Peter Novák (editors):
Proceedings of the 10th International Workshop on Computational Logic in Multi-Agent Systems 2009
IfI-09-08 Clausthal University of Technology, Germany, 2009
Niklas Skamriis Boss & Andreas Schmidt:
Implementing a Multi-Agent System
Special Course Report - Supervisor: Jørgen Villadsen
DTU Informatics 2009