基于大模型辅助的云边协同工作流调度算法

黎广镕; 李广军; 尚晶; 吴文泰; 王泽平; 龙赛琴

doi:10.12263/DZXB.20250494

您当前的位置：

首页 >

文章列表页 >

基于大模型辅助的云边协同工作流调度算法

大模型与互联网 | 更新时间：2025-12-27

- 基于大模型辅助的云边协同工作流调度算法
- Large Language Model-Assisted Cloud-Edge Collaborative Workflow Scheduling Algorithm
- 电子学报 2025年53卷第9期页码：3060-3077
- 作者机构：
  
  1.暨南大学信息科学技术学院，广东广州 510632
  2.中移动信息技术有限公司，北京 100033
- 作者简介：
  
  [ "黎广镕男，2004年2月出生于广东省廉江市.现为暨南大学信息科学技术学院软件工程专业本科生.E-mail: lgr2172@stu2022.jnu.edu.cn" ]
  [ "李广军男，2002年7月出生于贵州省遵义市.现为暨南大学人工智能专业硕士研究生.主要研究方向为大语言模型、机器学习系统.E-mail: lniox_1103@163.com" ]
  [ "尚晶女，1978年1月出生于河北省沧州市.博士，中国移动信息技术中心高级工程师，中国移动首席专家.主要研究方向为大数据、云计算、数据库. E-mail: shangjing@chinamobile.com" ]
  [ "吴文泰男，1992年11月出生于广东省广州市.英国华威大学计算机科学博士、暨南大学信息科学技术学院副教授.主要研究方向为并行与分布式计算、分布式机器学习以及节能计算. E-mail: wentaiwu@jnu.edu.cn" ]
  [ "王泽平男，1999年4月出生于贵州省凯里市.现为暨南大学信息科学技术学院博士研究生.主要研究方向为云计算、算力网络、人工智能和算力拍卖.E-mail: WangZP@stu2025.jnu.edu.cn" ]
  [ "龙赛琴女，1986年8月出生于湖南省益阳市.华南理工大学计算机应用技术博士、中国计算机学会成员、暨南大学信息科学技术学院教授.主要研究方向为云计算、云存储、并行和分布式系统、文件系统以及计算机系统架构.E-mail: saiqinlong@jnu.edu.cn" ]
- 基金信息：
  
  国家自然科学基金(U23B2027);广东基础与应用基础研究基金(2024A1515010214)
- DOI：10.12263/DZXB.20250494
  中图分类号： TP311.1;
- 收稿：2025-06-08，
  
  录用：2025-09-16，
  
  纸质出版：2025-09-25
- 稿件说明：
移动端阅览
黎广镕, 李广军, 尚晶, 等. 基于大模型辅助的云边协同工作流调度算法[J]. 电子学报, 2025, 53(09): 3060-3077.

LI Guang-rong, LI Guang-jun, SHANG Jing, et al. Large Language Model-Assisted Cloud-Edge Collaborative Workflow Scheduling Algorithm[J]. Acta Electronica Sinica, 2025, 53(09): 3060-3077.
黎广镕, 李广军, 尚晶, 等. 基于大模型辅助的云边协同工作流调度算法[J]. 电子学报, 2025, 53(09): 3060-3077. DOI：10.12263/DZXB.20250494

LI Guang-rong, LI Guang-jun, SHANG Jing, et al. Large Language Model-Assisted Cloud-Edge Collaborative Workflow Scheduling Algorithm[J]. Acta Electronica Sinica, 2025, 53(09): 3060-3077. DOI：10.12263/DZXB.20250494

摘要

工作流在云边协同环境中执行可以减少云与终端设备之间的数据传输时延.由于云计算节点、边缘设备在计算能力、存储资源及通信延迟等方面存在显著差异，加之边缘服务器计算资源受负载压力、性能退化等因素影响具有动态性，同时工作流应用内部复杂的拓扑依赖关系进一步增加了调度约束条件，使得该场景下的工作流调度问题被证明为NP-hard问题.针对上述问题，本文提出了基于大模型辅助的云边协同工作流调度算法（Large Language Model-Assisted Cloud-Edge Collaborative Workflow Scheduling Algorithm，LAWS）.该算法通过知识图谱结构化表征推理过程的思维链（Chain-of-Thought，CoT），将调度问题分解成多个子问题，并提取出子知识图谱作为子问题的思维链引导大模型协同推理调度决策.实验结果表明，与传统算法相比，该算法使得工作流执行时延降低3%~83%，计算能耗降低2.4%~66.0%.

Abstract

Executing workflows in cloud-edge collaborative environments can reduce data transmission latency between the cloud and terminal devices. Significant differences exist between cloud computing nodes and edge devices in terms of computational capability

storage resources

and communication latency. Furthermore

the computational resources of edge servers exhibit dynamicity due to factors like workload pressure and performance degradation. The complex topological dependencies within workflow applications introduce additional scheduling constraints. These combined factors render the workflow scheduling problem in this context NP-hard. To address these challenges

this paper proposes large language model-assisted cloud-edge collaborative workflow scheduling algorithm (LAWS). The algorithm employs a knowledge graph to structurally represent the chain-of-thought (CoT) reasoning process. It decomposes the scheduling problem into multiple sub-problems and extracts sub-knowledge graphs to serve as chain-of-thought guides for the large model

facilitating collaborative reasoning for scheduling decisions. Experimental results demonstrate that compared with traditional algorithms

the proposed algorithm achieves a reduction in workflow execution latency of 3% to 83% and a decrease in computational energy consumption of 2.4% to 66.0%.

关键词

Keywords

references

WANG T , LU Y C , WANG J H , et al . EIHDP: Edge-intelligent hierarchical dynamic pricing based on cloud-edge-client collaboration for IoT systems [J ] . IEEE Transactions on Computers , 2021 , 70 ( 8 ): 1285 - 1298 .

许悦玥 , 刘博文 , 田臣 , 等 . 基于联盟链的可靠边缘计算任务卸载方法 [J ] . 电子学报 , 2024 , 52 ( 1 ): 232 - 243 .

XU Y Y , LIU B W , TIAN C , et al . Task unloading method for reliable edge computing based on alliance chain [J ] . Acta Electronica Sinica , 2024 , 52 ( 1 ): 232 - 243 . (in Chinese)

丁婧伊 , 金嘉晖 , 杨丰赫 , 等 . 基于云边协作的工业互联网排产方法: 以钢铁热轧生产为例 [J ] . 电子学报 , 2024 , 52 ( 9 ): 2988 - 2999 .

DING J Y , JIN J H , YANG F H , et al . Industrial Internet scheduling method based on cloud-edge collaboration: A case study of steel hot rolling [J ] . Acta Electronica Sinica , 2024 , 52 ( 9 ): 2988 - 2999 . (in Chinese)

MA X J , XU H H , GAO H H , et al . Real-time multiple-workflow scheduling in cloud environments [J ] . IEEE Transactions on Network and Service Management , 2021 , 18 ( 4 ): 4002 - 4018 .

SENJAB K , ABBAS S , AHMED N , et al . A survey of Kubernetes scheduling algorithms [J ] . Journal of Cloud Computing , 2023 , 12 ( 1 ): 87 .

TANG X Y , CAO W B , TANG H Y , et al . Cost-efficient workflow scheduling algorithm for applications with deadline constraint on heterogeneous clouds [J ] . IEEE Transactions on Parallel and Distributed Systems , 2022 , 33 ( 9 ): 2079 - 2092 .

SUN Z X , ZHANG B Y , GU C L , et al . ET2FA: A hybrid heuristic algorithm for deadline-constrained workflow scheduling in cloud [J ] . IEEE Transactions on Services Computing , 2023 , 16 ( 3 ): 1807 - 1821 .

SHIN J , ARROYO D , TANTAWI A , et al . Cloud-native workflow scheduling using a hybrid priority rule, dynamic resource allocation, and dynamic task partition [C ] // Proceedings of the 2024 ACM Symposium on Cloud Computing . New York : ACM , 2024 : 830 - 846 .

LIAO H Y , LIU T Y , GUO J M , et al . Retrospecting available CPU resources: SMT-aware scheduling to prevent SLA violations in data centers [J ] . IEEE Transactions on Parallel and Distributed Systems , 2025 , 36 ( 1 ): 67 - 83 .

PALLEWATTA S , KOSTAKOS V , BUYYA R . Reliability-aware proactive placement of microservices-based IoT applications in fog computing environments [J ] . IEEE Transactions on Mobile Computing , 2024 , 23 ( 12 ): 11326 - 11341 .

XIA X W , QIU H X , XU X , et al . Multi-objective workflow scheduling based on genetic algorithm in cloud environment [J ] . Information Sciences , 2022 , 606 : 38 - 59 .

ZHOU J J , GAO L , RAO S J , et al . Scheduling constrained cloud workflow tasks via evolutionary multitasking optimization with adaptive knowledge transfer [J ] . IEEE Transactions on Services Computing , 2024 , 17 ( 6 ): 4254 - 4266 .

CHEN S W , YUAN Q F , LI J M , et al . Graph neural network aided deep reinforcement learning for microservice deployment in cooperative edge computing [J ] . IEEE Transactions on Services Computing , 2024 , 17 ( 6 ): 3742 - 3757 .

LIN L D , PAN L , LIU S J . SpotDAG: An RL-based algorithm for DAG workflow scheduling in heterogeneous cloud environments [J ] . IEEE Transactions on Services Computing , 2024 , 17 ( 5 ): 2904 - 2917 .

JAYANETTI A , HALGAMUGE S , BUYYA R . Multi-agent deep reinforcement learning framework for renewable energy-aware workflow scheduling on distributed cloud data centers [J ] . IEEE Transactions on Parallel and Distributed Systems , 2024 , 35 ( 4 ): 604 - 615 .

CHEN X , HU S X , YU C J , et al . Real-time offloading for dependent and parallel tasks in cloud-edge environments using deep reinforcement learning [J ] . IEEE Transactions on Parallel and Distributed Systems , 2024 , 35 ( 3 ): 391 - 404 .

JIAN Z L , XIE X S , FANG Y Z , et al . DRS: A deep reinforcement learning enhanced Kubernetes scheduler for microservice-based system [J ] . Software: Practice and Experience , 2024 , 54 ( 10 ): 2102 - 2126 .

WEN H , LI Y C , LIU G H , et al . AutoDroid: LLM-powered task automation in Android [C ] // Proceedings of the 30th Annual International Conference on Mobile Computing and Networking . New York : ACM , 2024 : 543 - 557 .

HE G L , DEMARTINI G , GADIRAJU U . Plan-then-execute: An empirical study of user trust and team performance when using LLM agents as a daily assistant [C ] // Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems . New York : ACM , 2025 : 1 - 22 .

FAKHOURY S , NAIK A , SAKKAS G , et al . LLM-based test-driven interactive code generation: User study and empirical evaluation [J ] . IEEE Transactions on Software Engineering , 2024 , 50 ( 9 ): 2254 - 2268 .

GU Q H . LLM-based code generation method for golang compiler testing [C ] // Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering . New York : ACM , 2023 : 2201 - 2203 .

ALI J , MAQSOOD T , KHALID N , et al . Communication and aging aware application mapping for multicore based edge computing servers [J ] . Cluster Computing , 2023 , 26 ( 1 ): 223 - 235 .

SCHROEDER B , GIBSON G A . A large-scale study of failures in high-performance computing systems [C ] // International Conference on Dependable Systems and Networks . Piscataway : IEEE , 2006 : 249 - 258 .

CHEN X J , JIA S B , XIANG Y . A review: Knowledge reasoning over knowledge graph [J ] . Expert Systems with Applications , 2020 , 141 : 112948 .

GAUTIER G , POLITO G , BARDENET R , et al . DPPy: Sampling DPPs with Python [EB/OL ] . ( 2019-08-12 )[ 2025-08-30 ] . https://arXiv.org/abs/1809.07258 https://arXiv.org/abs/1809.07258 .

HUANG L , YU W J , MA W T , et al . A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions [J ] . ACM Transactions on Information Systems , 2025 , 43 ( 2 ): 1 - 55 .

RAMPRASAD S , FERRACANE E , LIPTON Z . Analyzing LLM behavior in dialogue summarization: Unveiling circumstantial hallucination trends [C ] // Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics . Stroudsburg : ACL , 2024 : 12549 - 12561 .

DENIS C , HEBIRI M , ZAOUI A . Regression with reject option and application to kNN [EB/OL ] . ( 2021-03-05 )[ 2025-08-30 ] . https://arXiv.org/abs/2006.16597 https://arXiv.org/abs/2006.16597 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

一种基于子图转述的问题生成方法

基于多关系知识增强的开发者推荐算法

基于深度学习的实体关系联合抽取研究综述

基于知识图谱表示学习的谣言早期检测方法