Multi-Agent Coordination Adaptation via Structure-Guided Orchestration

2026-05-25

AI Summary

This work addresses the challenge of balancing structural stability and dynamic adaptability in large language model–driven multi-agent systems. The authors propose MACA, a novel framework that jointly models coordination structure and orchestration as a probabilistic inference problem. By leveraging a structure prior conditioned on task requirements and resource budgets, MACA enables fine-grained, controllable adaptive coordination through policy-guided orchestration. The approach maintains system stability while significantly enhancing efficiency, achieving an average performance gain of 8.42% across multiple benchmarks and reducing token consumption by 43.19%. This reduction effectively suppresses redundant agent interactions and accelerates convergence to high-efficiency execution pathways.

Abstract

As large language model (LLM)-based multi-agent systems scale to handle increasingly complex tasks, balancing structural stability and dynamic adaptability becomes increasingly challenging. Existing systems typically adopt either structure-centric methods, committing to structures determined upfront that limit fine-grained control, or orchestration-centric methods, adapting decisions dynamically while leaving coordination structure implicit and unstable. To address this challenge, we revisit multi-agent coordination from a probabilistic perspective, casting it as posterior inference over the joint distribution of structure and orchestration. We introduce MACA, an automated coordination framework that learns a task- and budget-conditioned structural prior over agent participation and interactions. This prior guides a policy-based orchestration as an approximation to posterior inference, enabling efficient solutions with fine-grained control. Across benchmarks, MACA outperforms adaptive multi-agent baselines by an average of 8.42% while using 43.19% fewer tokens. Further investigation reveals that joint adaptation of structure and orchestration suppresses redundant interactions, converging coordination toward task-effective execution.

Problem

• multi-agent coordination
• structural stability
• dynamic adaptability
• LLM-based systems
• coordination structure

Innovation

• multi-agent coordination
• structure-guided orchestration
• probabilistic inference
• LLM-based agents
• adaptive coordination