Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.25320 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914121899835392 |
|---|---|
| author | Wu, Jiaqi Zhao, Qinlao Chen, Zefeng Qin, Kai Zhao, Yifei Wang, Xueqian Yao, Yuhang |
| author_facet | Wu, Jiaqi Zhao, Qinlao Chen, Zefeng Qin, Kai Zhao, Yifei Wang, Xueqian Yao, Yuhang |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_25320 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GAP: Graph-Based Agent Planning with Parallel Tool Use and Reinforcement Learning Wu, Jiaqi Zhao, Qinlao Chen, Zefeng Qin, Kai Zhao, Yifei Wang, Xueqian Yao, Yuhang Artificial Intelligence Computation and Language 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. |
| title | GAP: Graph-Based Agent Planning with Parallel Tool Use and Reinforcement Learning |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2510.25320 |