Proceedings

Multi-Agent Intention Progression with Black-Box Agents

Michael Dann, Yuan Yao, Brian Logan, John Thangarajah

Download BibTeX Cite

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. Main Track. Pages 132-138.

https://doi.org/10.24963/ijcai.2021/19

Keywords

  • Agent-based and Multi-agent Systems: Engineering Methods, Platforms, Languages and Tools
  • Agent-based and Multi-agent Systems: Agent Theories and Models

Abstract

We propose a new approach to intention progression in multi-agent settings where other agents are effectively black boxes. That is, while their goals are known, the precise programs used to achieve these goals are not known. In our approach, agents use an abstraction of their own program called a partially-ordered goal-plan tree (pGPT) to schedule their intentions and predict the actions of other agents. We show how a pGPT can be derived from the program of a BDI agent, and present an approach based on Monte Carlo Tree Search (MCTS) for scheduling an agent's intentions using pGPTs. We evaluate our pGPT-based approach in cooperative, selfish and adversarial multi-agent settings, and show that it out-performs MCTS-based scheduling where agents assume that other agents have the same program as themselves.

Michael Dann, Yuan Yao, Brian Logan, and John Thangarajah. 2021. Multi-Agent Intention Progression with Black-Box Agents. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21. IJCAI, 132–138. https://doi.org/10.24963/ijcai.2021/19
Welcome Future Conferences AI Journal Division IJCAI Proceedings About IJCAI IJCAI Awards