多目标优化的SWIPT-MEC任务分级卸载架构与优化算法  

Multi-objective optimization-based hierarchical offloading architecture and optimization algorithm for SWIPT-MEC tasks

作　　者：张泽维 李陶深[1,2] 杨林峰 ZHANG Ze-wei;LI Tao-shen;YANG Lin-feng(School of Computer,Electronics and Information,Guangxi University,Nanning 530004,China;China-ASEAN International Join Laboratory of Integrate Transport,Nanning University,Nanning 530200,China)

机构地区：[1]广西大学计算机与电子信息学院,广西南宁530004 [2]南宁学院中国东盟综合交通国际联合实验室,广西南宁530200

出　　处：《广西大学学报（自然科学版）》2022年第2期487-497,共11页Journal of Guangxi University（Natural Science Edition）

基　　金：国家自然科学基金项目(61762010);广西科技计划项目(桂科AD20297125)。

摘　　要：为了解决单用户模式的MEC卸载策略未能有效地权衡移动设备在计算过程中的能耗与时耗之间关系的问题,针对无线携能通信网络(SWIPT),构建一种单用户单核服务器的多任务分级卸载架构与多目标数学模型,在卸载操作过程中,根据计算任务所需的时延和能耗,为任务选择合适的卸载方式,解决某时段内任务密集场景下的任务卸载问题,提升设备整体执行效率。然后,提出一种基于多目标优化改进的强度帕累托进化算法(SPEA2)的任务分级卸载架构与优化算法,最大限度地权衡MEC卸载的低能耗与低时耗。仿真实验验证了多目标MEC任务分级卸载求解策略可有效降低设备的能量消耗率和时间消耗率。In order to solve the problem that the single-user mode mobile edge computing(MEC)offloading strategy fails to effectively balance the relationship between energy consumption and time consumption of mobile devices in the calculation process,a multitask hierarchical offloading architecture and multi-objective mathematical model of single user and single core server are constructed.During the offloading operation,the model can select an appropriate offloading method for the task according to the time delay and energy consumption required by the calculation task,solve the problem of task offloading in the task intensive scenario in a certain period of time,and improve the overall execution efficiency of the equipment.Then,a hierarchical task offloading architecture and optimization algorithm based on multi-objective optimization improved strength Pareto evolutionary algorithm(SPEA2)are proposed for simultaneous wireless information and power transfer(SWIPT).Then,a task hierarchical unloading architecture and optimization algorithm based on multi-objective optimization improved intensity Pareto evolutionary algorithm(SPEA2)are proposed to better balance the low energy consumption and low time consumption of MEC offloading.The simulation experiment verifies that the multi-objective MEC task hierarchical offloading solution strategy can effectively reduce the energy consumption rate and time consumption rate of the equipment.

关 键 词：移动边缘计算 无线携能通信 多目标优化 改进的强度帕累托进化算法 分级卸载 

分 类 号：TP400[自动化与计算机技术]