用轻量模型统一调度多智能体与工具，实现并行高效协作
Small Model as Master Orchestrator: Learning Unified Agent-Tool Orchestration with Parallel Subtask Decomposition
- 将智能体和工具统一为可学习的动作空间，支持并行任务分解
- 在多个基准上表现优异，且对未见过的模型组合仍具泛化能力
- 适合需要高效多智能体协同的复杂任务系统设计
多智能体系统在解决复杂问题时具有明显优势，但现有编排方法多依赖静态流程或串行调度，且受限于智能体与工具间异构接口协议，导致系统复杂度高、扩展性差。为此，本文提出 Agent-as-Tool 统一并行编排范式，将智能体和工具抽象为标准化、可学习的动作空间，通过协议归一化与显式状态反馈实现统一管理。在此基础上，训练轻量级编排器 ParaManager，解耦规划决策与子任务求解，支持状态感知的并行子任务分解、委派与异步执行。采用两阶段训练流程：先通过带恢复机制的监督微调（SFT）提升鲁棒性，再结合强化学习（RL）优化任务成功率、协议合规性、多样性与推理效率的平衡。实验表明，ParaManager 在多个基准上表现强劲，并在未见模型池下展现出强泛化能力。
原文摘要 · Abstract (English)
Multi-agent systems (MAS) demonstrate clear advantages in tackling complex problems by coordinating diverse agents and external tools. However, most existing orchestration methods rely on static workflows or serial agent scheduling, and are further constrained by heterogeneous interface protocols between tools and agents. This leads to high system complexity and poor extensibility. To mitigate these issues, we propose Agent-as-Tool, a unified parallel orchestration paradigm that abstracts both agents and tools into a standardized, learnable action space with protocol normalization and explicit state feedback. Building on this paradigm, we train a lightweight orchestrator, ParaManager, which decouples planning decisions from subtask solving, enabling state-aware parallel subtask decomposition, delegation, and asynchronous execution. For training, we adopt a two-stage ParaManager training pipeline. It improves robustness by incorporating supervised fine-tuning (SFT) trajectories equipped with recovery mechanisms, and further applies reinforcement learning (RL) to achieve an optimal balance among task success, protocol compliance, diversity, and reasoning efficiency. Experiments show that ParaManager achieves strong performance across multiple benchmarks and exhibits robust generalization under unseen model pools.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。