XAgents: A Structured Rule-Guided Multi-Agent Cooperative System via Multipolar Task Graph Modeling

This paper proposes XAgents, a structured rule-guided multi-agent cooperative system that uses multipolar task graph modeling to coordinate complex multi-agent workflows. The system ensures reliable and interpretable agent collaboration through explicit rule-based guidance.

Submitted to IEEE Transactions on Emerging Topics in Computational Intelligence (IEEE TETCI). Currently under review (在审).

First author.