| 引用本文： |
-
张晓莉,雷雨声,刘夏茜,等.基于深度强化学习的工业SDN网络切片资源分配[J].电讯技术，2025，（8）：1221 - 1230. [点击复制]
- ZHANG Xiaoli,LEI Yusheng,LIU Xiaxi,et al.Industrial SDN Network Slicing Resource Allocation Based on Deep Reinforcement Learning[J].，2025，（8）：1221 - 1230. [点击复制]
|
|
| 摘要: |
| 针对工业物联网中业务需求多样性和服务质量（Quality of Service，QoS）要求差异性导致的网络资源利用低问题，提出一种基于深度强化学习的网络切片资源分配策略。该策略运用深度强化学习优化网络切片资源分配的准入控制，通过智能体在特定时间窗口内处理资源请求，并根据不同网络切片的QoS要求及请求准入结果进行资源的动态分配。实验结果表明，所提策略相比基准算法在提高网络收益、资源利用率和接收率方面分别提升了8.33%、9.84%和8.57%。该策略能够在保证服务质量的同时提高整个网络的效率和性能。 |
| 关键词: 工业物联网（IIOT） 软件定义网络；网络切片；资源分配；准入控制；深度强化学习 |
| DOI：10.20079/j.issn.1001-893x.240520002 |
|
| 基金项目:国家自然科学基金资助项目(U19B2015) |
|
| Industrial SDN Network Slicing Resource Allocation Based on Deep Reinforcement Learning |
| ZHANG Xiaoli,LEI Yusheng,LIU Xiaxi,WANG Bin |
| (School of Communication and Information Engineering,Xi’an University of Science and Technology,Xi’an 710600,China) |
| Abstract: |
| To address the issue of low network resource utilization caused by the diversity of business requirements and differences in quality of service(QoS) demands in industrial Interhot of Thiugs(IoT),a network slicing resource allocation strategy based on deep reinforcement learning is proposed.This strategy uses deep reinforcement learning to optimize the admission control of network slicing resource allocation.It processes resource requests within a specific time window through an agent and dynamically allocates resources based on the QoS requirements of different network slices and the admission results of the requests.Experimental results show that the proposed strategy improves network revenue,resource utilization,and acceptance rate by 8.33%,9.84%,and 8.57%,respectively,compared with the baseline algorithm.The strategy can improve the efficiency and performance of the entire network while ensuring service quality. |
| Key words: industrinal IoT(IIOT) software-defined networking network slicing resource allocation admission control deep reinforcement learning |
《电讯技术》编辑部 