让大模型任务规划更高效，通过图结构并行执行工具。
GAP: Graph-Based Agent Planning with Parallel Tool Use and Reinforcement Learning
- 用图结构建模任务依赖关系，智能决定哪些工具可并行执行。
- 在多跳问答数据集上，工具调用效率提升显著，准确率远超传统方法。
- 适合需要复杂推理与工具协同的自动化系统开发者使用。
由大语言模型驱动的自主代理在复杂任务求解中的工具操作能力已展现惊人潜力。然而，现有范式如 ReAct 依赖于串行推理与执行，无法利用独立子任务间的固有并行性，导致工具利用率低下且多步推理性能不佳。本文提出图基代理规划（GAP），通过图结构显式建模任务间依赖关系，实现自适应的并行与串行工具执行。该方法训练代理基础模型将复杂任务分解为依赖感知的子任务图，自主判断哪些工具可并行执行、哪些需顺序依赖。这种依赖感知调度在执行效率和任务准确性上均取得显著提升。为训练 GAP，我们基于多跳问答（MHQA）基准构建了高质量的图式规划轨迹数据集，并采用两阶段训练策略：先在筛选数据集上进行监督微调（SFT），再在具有最大工具推理价值的战略采样查询上进行基于正确性的强化学习（RL）。实验结果表明，GAP 在 MHQA 数据集上显著优于传统 ReAct 基线，尤其在多步检索任务中表现突出，同时通过智能并行化大幅提高工具调用效率。
原文摘要 · Abstract (English)
Autonomous agents powered by large language models (LLMs) have shown impressive capabilities in tool manipulation for complex task-solving. However, existing paradigms such as ReAct rely on sequential reasoning and execution, failing to exploit the inherent parallelism among independent sub-tasks. This sequential bottleneck leads to inefficient tool utilization and suboptimal performance in multi-step reasoning scenarios. We introduce Graph-based Agent Planning (GAP), a novel framework that explicitly models inter-task dependencies through graph-based planning to enable adaptive parallel and serial tool execution. Our approach trains agent foundation models to decompose complex tasks into dependency-aware sub-task graphs, autonomously determining which tools can be executed in parallel and which must follow sequential dependencies. This dependency-aware orchestration achieves substantial improvements in both execution efficiency and task accuracy. To train GAP, we construct a high-quality dataset of graph-based planning traces derived from the Multi-Hop Question Answering (MHQA) benchmark. We employ a two-stage training strategy: supervised fine-tuning (SFT) on the curated dataset, followed by reinforcement learning (RL) with a correctness-based reward function on strategically sampled queries where tool-based reasoning provides maximum value. Experimental results on MHQA datasets demonstrate that GAP significantly outperforms traditional ReAct baselines, particularly on multi-step retrieval tasks, while achieving dramatic improvements in tool invocation efficiency through intelligent parallelization. The project page is available at: https://github.com/WJQ7777/Graph-Agent-Planning.
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。