Argonauts: a working system for motivated cooperative agents

作者：Hoelzgen Daniel; Vengels Thomas; Kruempelmann Patrick; Thimm Matthias^*; Kern I**erner Gabriele

来源：Annals of Mathematics and Artificial Intelligence, 2011, 61(4): 309-332.

DOI：10.1007/s10472-011-9226-1

• BDI
• Multi-agent systems
• Motivation
• Answer set programming

This paper presents the Argonauts multi-agent framework which was developed as part of a one year student project at Technische Universitat Dortmund. The Argonauts framework builds on a BDI approach to model rational agents that act cooperatively in a dynamic and indeterministically changing environment. However, our agent model extends the traditional BDI approach in several aspects, most notably by incorporating motivation into the agent's goal selection mechanism. The framework has been applied by the Argonauts team in the 2010 version of the annual multi-agent programming contest organized by Technische Universitat Clausthal. In this paper, we present a high-level specification and analysis of the actual system used for solving the given scenario. We do this by applying the GAIA methodology, a high-level and iterative approach to model communication and roles in multi-agent scenarios. We further describe the technical details and insights gained during our participation in the multi-agent programming contest.

• 出版日期2011-4