1.South-Central Minzu University a. College of Electronic and Information Engineering
b.Hubei Key Laboratory of Intelligent Wireless Communications,Wuhan 430074,China
2.Hubei Engineering Research Center of Intelligent Internet of Things Technology,Wuhan 430074,China
Show less
文章历史+
Received
Accepted
Published
2024-04-04
Issue Date
2025-10-27
PDF (2413K)
摘要
针对物联网(Internet of Things,IoT)场景下具有差异化资源需求的网络服务资源分配问题,提出了一种将网络切片技术(Network Slicing, NS)与确定性策略梯度(Deep Deterministic Policy Gradient,DDPG)相结合的动态切片资源分配算法(Dynamic Slicing Resource Allocation,DSRA).该算法根据切片上不同设备的资源需求,动态分配虚拟化的无线接入网资源,以满足设备资源需求并最小化系统总成本.仿真实验对比分析了所提出的算法与四种基线算法在服务质量(Quality of Service,QoS)满意率和系统总成本的表现.仿真结果表明:在具备多种资源与大量设备的场景中,所提出的算法与基线算法相比,能够显著提高设备的QoS水平,并降低系统的总成本.
Abstract
A Dynamic Slicing Resource Allocation algorithm (DSRA) based on Network Slicing (NS) and Deep Deterministic Policy Gradient (DDPG) is proposed to address the resource allocation problem of network services with differentiated resource demands in the Internet of Things (IoT). The algorithm dynamically allocates virtual radio access network resources based on the resource requirements of different devices on the slice, to meet device resource requirements and minimize the total system costs. The simulation experiment compared and analyzed the performance of the proposed algorithm with four baseline algorithms in terms of quality of service (QoS) satisfaction rate and total system cost. The simulation results show that in scenarios with multiple resources and a large number of devices, the proposed algorithm can significantly improve the QoS level of devices and reduce the total cost of the system compared to baseline algorithms.
如今无线网络逐渐发展成为规模庞大的异构物联网(Internet of Things,IoT),新兴的IoT应用使网络中连接的设备数量呈现指数级增长[1],应用场景日益多样化.传统网络只能在一个物理网络基础设施中实现单一的网络资源配置,无法为运行在同一物理网络上的不同应用场景提供差异化的网络需求.为应对物联网设备数量的大规模增长,满足不同类型的设备或应用提出的服务质量需求,下一代移动网络联盟在2015年首次引入了网络切片技术的概念[2].网络切片技术将物理网络中的实际资源虚拟化,在同一物理基础设施中切分出多个包含网络功能、网络资源和特定配置的虚拟网络,每个虚拟网络能依据其所包含的不同业务按需提供网络资源,形成自己独特的容量、带宽、延迟和安全等特性[3-4].
目前,资源分配方案是网络切片领域中的研究热点,通过在切片之间动态调整网络中的资源,能提高网络性能和资源利用率,满足用户的QoS需求[5].文献[6]和文献[7]将物联网中资源分配问题建模为混合整数线性规划(Mixed Integer Linear Programming,MILP)问题.但由于资源分配问题的状态空间和动作空间都是非常复杂的,该优化问题通常是一个NP-hard问题,上述基于优化方案只能获得近似最优解,而文献[8-11]则采用启发式算法解决NP-hard问题.文献[8]采用MILP公式描述切片准入控制方案,设计了基于贪婪启发式方法来求解资源分配问题.文献[9]采用启发式方法计算多项式时间内每个切片的可行和次优资源分配.但上述文献侧重于最大化切片接受度和充分利用物理资源,忽视了系统节能问题.最小化总能源成本至关重要,且对最大化服务提供商的净利润起到积极影响.文献[10]针对如何最小化切片资源分配过程中产生的总能耗,提出了基于整数线性规划(Integer Linear Programming,ILP)和启发式算法的解决方案.文献[11]为最大化满足切片资源请求的总回报并最小化切片资源再分配的总惩罚,提出一个贪婪启发式算法进行求解.上述系统本质上都是静态或准静态的,对于5G网络中设备接入的高动态特性,这些资源分配方法在实际应用时需要根据情况的变化而重新求解,从而消耗大量的算力.面对动态的切片资源分配问题,基于机器学习的方法更具优势.
WIJETHILAKAS, LIYANAGEM. Survey on network slicing for Internet of Things realization in 5G networks[J]. IEEE Communications Surveys & Tutorials, 2021, 23(2): 957-994.
[2]
PORAMBAGEP, OKWUIBEJ, LIYANAGEM, et al. Survey on multi-access edge computing for Internet of Things realization[J]. IEEE Communications Surveys & Tutorials, 2018, 20(4): 2961-2991.
[3]
CAOH, DUJ, ZHAOH, et al. Toward tailored resource allocation of slices in 6G networks with softwarization and virtualization[J]. IEEE Internet of Things Journal, 2022, 9(9): 6623-6637.
HARUTYUNYAND, FEDRIZZIR, SHAHRIARN, et al. Orchestrating end-to-end slices in 5G networks[C]//2019 15th International Conference on Network and Service Management (CNSM). Halifax:IEEE, 2019: 1-9.
[6]
YINB, TANGJ, WENTM. Maximizing the connectivity of wireless network slicing enabled industrial Internet-of-things[C]//2021 IEEE Global Communications Conference (GLOBECOM). Madrid: IEEE, 2021: 1-6.
DEBBABIF, TAKTAKR, JMALR, et al. Inter-slice B5G bandwidth resource allocation[C]//2022 IEEE 21st International Symposium on Network Computing and Applications (NCA). Boston:IEEE, 2022: 157-163.
[9]
CAOH, DUJ, ZHAOH, et al. Resource-ability assisted service function chain embedding and scheduling for 6G networks with virtualization[J]. IEEE Transactions on Vehicular Technology, 2021, 70(4): 3846-3859.
[10]
JANGI, SUH D, PACKS, et al. Joint optimization of service function placement and flow distribution for service function chaining[J]. IEEE Journal on Selected Areas in Communications, 2017, 35(11): 2532-2541.
[11]
BERAS, MEHTAN B. Network slicing in 5G edge networks with controlled slice redistributions[C]//2021 17th International Conference on Network and Service Management (CNSM). Izmir:IEEE, 2021: 118-124.
[12]
SHENX, GAOJ, WUW, et al. AI-assisted network-slicing based next-generation wireless networks[J]. IEEE Open Journal of Vehicular Technology, 2020, 1: 45-66.
[13]
LIR, WANGC, ZHAOZ, et al. The LSTM-based advantage actor-critic learning for resource management in network slicing with user mobility[J]. IEEE Communications Letters, 2020, 24(9): 2005-2009.
[14]
HUAY, LIR, ZHAOZ, et al. GAN-powered deep distributional reinforcement learning for resource management in network slicing[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(2): 334-349.
[15]
HAOM, YED, WANGS, et al. URLLC resource slicing and scheduling in 5G vehicular edge computing[C]//2021 IEEE 93rd Vehicular Technology Conference (VTC2021-Spring). Helsinki:IEEE, 2021: 1-5.
[16]
LILLICRAPT P, HUNTJ J, PRITZELA, et al. Continuous control with deep reinforcement learning[J]. arXiv:2015,1509.02971.