In fog computing networks, due to the limited service coverage of fog computing node(FCN), the moving users may be out of the coverage, which would cause the migration of execution results. Furthermore, energy consumption and time latency are generated and affects the revenue of users. To solve these problems, a generic three layers fog computing networks is considered, and the mobility of users is characterized by the sojourn time. To maximize the revenue of users and reduce system overhead, the offloading decisions, FCN selection and computation resource allocation are jointly optimized to reduce the probability of migration. The maximization problem is divided into two parts: task offloading and resource allocation. A FCN selection algorithm based on Gini coefficient (FSAGC) is proposed to get a task offloading strategy, and a distributed resource allocation algorithm based on genetic algorithm (DRAAGA) is implemented to solve the computing resource allocation problem. Simulations demonstrate that the proposed algorithms can achieve more revenue performance compared with other algorithms.
因此,与以往独立研究不同,本文在三层雾计算网络中,通过用户驻留时间分析用户的移动性,并通过联合考虑任务卸载策略、FCN选择和计算资源分配来减少任务迁移的概率,降低系统开销,进而最大化用户的总收益。本文研究的用户收益最大化问题可以转换为一个混合整数非线性规划问题,该问题是一个非确定性多项式难题(non-deterministic polynomial-hard problem,NP-hard problem),可分解为任务卸载和资源分配两部分,我们分别采用基于基尼系数的雾计算节点选择算法(FCN selection algorithm based on Gini coefficient,FSAGC)和基于遗传算法的分布式资源分配算法(distributed resource allocation algorithm based on genetic algorithm,DRAAGA)进行求解。
本文考虑正交频分多址接入(orthogonal frequency division multiple access,OFDMA)上行系统采用多接入的方式。假设在一个FCN的覆盖区域内,一个FCN能同时连接多个用户设备。即使设备在重叠覆盖区域内,每个用户设备也只能同时连接到一个FCN中。定义集合表示包含所有FCN选择的变量集合,即,表示用户设备选择卸载它的任务给第f个FCN即。选择策略必须满足如下限制
根据文献[18],基尼系数常用于经济学领域用来衡量收益差距的指标,我们受到基尼系数的启发,采用基尼系数解决FCN选择问题。本文每个用户设备的收益被定义为一个选择函数,不同的用户设备对系统总收益有不同贡献。采用选择函数获得,这里被选中的用户设备可以贡献主要收益,收益包括能量、时间和迁移量。因此基于基尼系数的雾计算节点选择算法(FCN selection algorithm based on Gini coefficient, FSAGC),详见算法1。
当任务卸载决策和FCN选择完成后,(23)式的求解问题转化为对资源进行分配的求解问题,常见的求解方法有回溯法[15]、动态规划法算法[15]等,但是这些算法需要较多的迭代次数,时间复杂度和空间复杂度都较高。为了减少迭代次数和时间、空间复杂度,本文提出了一种基于遗传算法的分布式资源分配算法(distributed resource allocation algorithm based on genetic algorithm, DRAAGA),如算法2。该算法受达尔文进化论的启发,是一种启发式搜索方法,用于逼近最佳解,算法过程体现了自然选择,它选择最合适的个体进行后代遗传,该算法广泛应用于机器学习、组合优化和智能计算中。
LIUX N, ZHANGH J, LONGK P, et al. Energy efficient user association, resource allocation and caching deployment in fog radio access networks[J]. IEEE Transactions on Vehicular Technology, 2022, 71(2): 1846-1856. DOI: 10.1109/TVT.2021.3131720 .
[2]
LIUL, CHANGZ, GUOX. Socially aware dynamic computation offloading scheme for fog computing system with energy harvesting devices[J]. IEEE Internet of Things Journal, 2018, 5(3): 1869-1879. DOI:10.1109/JIOT.2018.2816682 .
[3]
SHAH MANSOURIH, WONGV W S. Hierarchical fog-cloud computing for IoT systems: A computation offloading game[J]. IEEE Internet of Things Journal, 2018, 5(4): 3246-3257. DOI: 10.1109/JIOT.2017.2788802 .
[4]
ZHANGG W, SHENF, YANGY,et al. Fair task offloading among fog nodes in fog computing networks[C]// 2018 IEEE International Conference on Communications. New York: IEEE Press,2018: 1-6. DOI: 10.1109/ICC.2018.8422316 .
[5]
BOZORGCHENANIA, TARCHID, CORAZZAG E. Centralized and distributed architectures for energy and delay efficient fog network-based edge computing services[J]. IEEE Transactions on Green Communications and Networking, 2019, 3(1): 250-263. DOI: 10.1109/TGCN.2018.2885443 .
[6]
YOUSEFPOURA, ISHIGAKIG, GOURR, et al. On reducing IoT service delay via fog offloading [J]. IEEE Internet of Things Journal, 2018, 5(2): 998-1010. DOI: 10.1109/JIOT.2017.2788802 .
[7]
XIONGZ H, FENGS H, WANGW B, et al. Cloud/fog computing resource management and pricing for blockchain networks [J]. IEEE Internet of Things Journal, 2019, 6(3): 4585-4600. DOI: 10.1109/JIOT.2018.2871706 .
[8]
LIUL Q, CHANGZ, GUOX J, et al. Multiobjective optimization for computation offloading in fog computing[J]. IEEE Internet of Things Journal, 2018, 5(1): 283-294. DOI: 10.1109/JIOT.2017.2780236 .
[9]
CHITIF, FANTACCIR, PICANOB. A matching theory framework for tasks offloading in fog computing for IoT systems [J]. IEEE Internet of Things Journal, 2018, 5(6): 5089-5096. DOI: 10.1109/JIOT.2018.2871251 .
[10]
LIG L, WUJ, LIJ H, et al. Service popularity-based smart resources partitioning for fog computing-enabled industrial Internet of Things [J]. IEEE Transactions on Industrial Informatics, 2018, 14(10): 4702-4711. DOI: 10.1109/TII.2018.2845844 .
[11]
RAHMANA, JINJ, CRICENTIA, et al. Motion and connectivity aware offloading in cloud robotics via genetic algorithm [C]// 2017 IEEE Global Communications Conference. New York: IEEE Press,2017: 1-6. DOI:10.1109/GLOCOM.2017.8255040 .
SHIY, CHENS Z, XUX. MAGA: A mobility-aware computation offloading decision for distributed mobile cloud computing[J]. IEEE Internet of Things Journal, 2018, 5(1): 164-174. DOI: 10.1109/JIOT.2017.2776252 .
[14]
SUNY Z, GUOG, ZHANGS Q, et al. A cluster-based energy-efficient resource management scheme with QoS requirement for ultra-dense networks[J]. IEEE Access, 2020, 8(4):182412-182421. DOI:10.1109/ACCESS.2020.3028942 .
[15]
GAOS, PENGZ, XIAOB, et al. Detection and mitigation of DoS attacks in software defined networks [J]. IEEE/ACM Transactions on Networking, 2020, 28(3): 1419-1433. DOI:10.1109/TNET.2020.2983976 .
[16]
ZHANGY M, LANX L, RENJ, et al. Efficient computing resource sharing for mobile edge-cloud computing networks[J]. IEEE/ACM Transactions on Networking, 2020, 28(3):1227-1240. DOI:10.1109/TNET.2020.2979807 .
[17]
ZENGJ, SUNJ Y, WUB W, et al. Mobile edge communications, computing, and caching(MEC3) technology in the maritime communication network[J]. China Communications, 2020, 17(5):223-234. DOI:10.23919/JCC.2020.05.017 .
[18]
WENY G, ZHANGW W, LUOH Y. Energy-optimal mobile application execution: Taming resource-poor mobile devices with cloud clones[C]// 2012 IEEE INFOCOM. Orlando: IEEE Press, 2012: 1-6. DOI: 10.1109/INFCOM.2012.6195685 .
[19]
MAHBASA J, ZHUH L, WANGJ Z. Impact of small cells overlapping on mobility management[J]. IEEE Transactions on Wireless Communications, 2019, 18(2): 1054-1068. DOI:10.1109/TWC.2018.2889465 .