---
格式版本: 2
标题: "A Theoretical Framework for Parallel Lifelong MAPF Using Group Decentralized Planning"
原文链接: "https://arxiv.org/abs/2608.17928"
发布日期: "2026-08-18"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Tue, 18 Aug 2026 15:48:18 UTC (5,432 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T15:46:10+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T15:43:02+08:00"
入库时间: "2026-08-20T07:46:10.915Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=throughput&searchtype=all"
匹配关键词:
  - "throughput"
  - "AI"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 10
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "论文主题为多智能体路径规划，与超节点/AI Rack等完全无关，仅命中throughput关键词。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T15:51:42+08:00"
AI主题相关性: 0
AI来源权威性: 5
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 5
AI摘要: "针对终身多智能体路径规划问题，作者提出分组去中心化RHCR（GD-RHCR）框架，在理论上证明滚动时域冲突消解（RHCR）在折扣MDP模型中具有近似最优保证，并通过基于传递通信的分组并行规划，在不同地图上以更低规划成本扩展到更高智能体数量。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:22:37.626Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.17928"
---

## Computer Science > Multiagent Systems

## Title:A Theoretical Framework for Parallel Lifelong MAPF Using Group Decentralized Planning

Authors:[Alex DeWeese](https://arxiv.org/search/cs?searchtype=author&query=DeWeese,+A), [Jiaoyang Li](https://arxiv.org/search/cs?searchtype=author&query=Li,+J), [Guannan Qu](https://arxiv.org/search/cs?searchtype=author&query=Qu,+G)

[View PDF](https://arxiv.org/pdf/2608.17928) [HTML (experimental)](https://arxiv.org/html/2608.17928v1)

> Abstract:In the Lifelong Multi-Agent Path Finding (L-MAPF) problem, agents must repeatedly move from one destination to another while avoiding obstacles and inter-agent collisions. Widely regarded as one of the highest-performing solutions to this problem is the Rolling-Horizon Collision Resolution (RHCR) framework. However, commensurate with its quality solutions, it incurs a computational cost that limits its applicability to even modest agent counts. In this paper, leveraging theoretical methods from the Locally Interdependent Multi-Agent MDP literature, we first theoretically prove the near-optimality of RHCR in a discounted MDP formulation of the L-MAPF problem. Then, we leverage these results to naturally motivate an extended framework called Group Decentralized RHCR (GD-RHCR) which incorporates a group decentralized structure that partitions agents based on a transitive communication scheme and plans for each partition of agents in parallel. We show that both RHCR and GD-RHCR achieve similar exponentially close to optimal guarantees, establishing a theoretical duality between the time based restrictions performed by vanilla RHCR and the additional space based partitioning performed by GD-RHCR. Lastly, we show that across varying maps, GD-RHCR is able to attain high throughput that scales into higher agent counts while maintaining a significantly lower per plan cost.

| Subjects: | Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Robotics (cs.RO) |
| --- | --- |
| Cite as: | [arXiv:2608.17928](https://arxiv.org/abs/2608.17928) \[cs.MA\] |
|  | (or [arXiv:2608.17928v1](https://arxiv.org/abs/2608.17928v1) \[cs.MA\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.17928](https://doi.org/10.48550/arXiv.2608.17928) |

## Submission history

From: Alex DeWeese \[[view email](https://arxiv.org/show-email/84a608c2/2608.17928)\]  
**\[v1\]** Tue, 18 Aug 2026 15:48:18 UTC (5,432 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.17928) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))