Gated Coordination for Efficient Multi-Agent Collaboration in Minecraft Game

2026-04-20

AI Summary

Existing open-world multi-agent approaches often trigger communication automatically in response to local anomalies, which can introduce coordination noise, disrupt local execution, and lead to redundant messaging. This work proposes a partitioned information architecture that decouples private execution from shared coordination states and, for the first time, formulates communication as a cost-sensitive, selective decision-making process. By integrating event-triggered working memory with a system-verification-based gating mechanism, the method activates communication only when necessary, ensuring efficiency. Evaluated on long-horizon construction tasks in Minecraft, the approach significantly outperforms strong communication baselines, achieving higher blueprint completion quality and longer execution chains while enhancing local self-recovery capabilities, reducing ineffective communication, and improving the utility of public interactions.

Technology Category

• Multiagent Systems: Agent Communication
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

In long-horizon open-world multi-agent systems, existing methods often treat local anomalies as automatic triggers for communication. This default design introduces coordination noise, interrupts local execution, and overuses public interaction in cases that could be resolved locally. To address this issue, we propose a partitioned information architecture for MLLM agents that explicitly separates private execution states from public coordination states. Building on this design, we introduce two key mechanisms. First, we develop an event-triggered working memory based on system-verified outcomes to maintain compact and low-noise local state representations. Second, we propose a cost-sensitive gated escalation mechanism that determines whether cross-region communication should be initiated by jointly considering node criticality, local recovery cost, and downstream task impact. In this way, communication is transformed from a default reaction into a selective decision. Experiments conducted on long-term construction tasks in open environments demonstrate that, compared to baseline models based on strong communication and planned structures, the introduction of gated communication and a partitioned information architecture results in superior performance in terms of blueprint completion quality and execution chain length. It also improves local self-recovery, reduces ineffective escalations, and increases the utility of public communication.

Problem

• multi-agent collaboration
• coordination noise
• communication overuse
• local anomaly handling
• open-world systems

Innovation

• gated communication
• partitioned information architecture
• event-triggered working memory
• cost-sensitive escalation
• multi-agent coordination