多核服务器边缘计算系统中任务卸载调度和功率分配的研究
    点此下载全文
引用本文:凌雪延,王鸿,宋荣方.多核服务器边缘计算系统中任务卸载调度和功率分配的研究[J].南京邮电大学学报:自然科学版,2020,40(2):81~88
摘要点击次数: 232
全文下载次数: 493
作者单位
凌雪延 南京邮电大学 通信与信息工程学院江苏南京210003 
王鸿 南京邮电大学 通信与信息工程学院江苏南京210003 
宋荣方 南京邮电大学 通信与信息工程学院江苏南京210003南京邮电大学 江苏省通信与网络技术工程研究中心江苏南京210003 
中文摘要:在单用户多核服务器的移动边缘计算系统中,对多个独立计算任务的卸载调度策略和功率分配问题进行了研究。首先,基于混合流水车间调度模型对系统任务的卸载调度进行了建模,获得了系统时延的计算表达式,并在此基础上对系统能耗进行了建模;其次,利用混合编码方式的遗传算法,对系统时延和能耗的加权和最小化问题进行了求解,确定了最优的任务卸载调度策略和系统时延与能耗间的权衡关系。通过仿真分析,获得了最优的任务卸载甘特图;揭示了系统时延随任务数量线性增长;与随机任务卸载策略相比,同样条件下,文中所提出的卸载策略具有较小的系统时延;同时,可得到系统能耗和系统时延成反比的关系,在不增加系统时延的条件下找到了有效的节能方式。
中文关键词:移动边缘计算  多核服务器  任务卸载调度策略  功率控制  混合流水车间调度  遗传算法
 
Joint task offloading scheduling and transmit power allocation for mobile edge computing systems with multi core server
Abstract:Joint task offloading scheduling and transmit power allocation for multiple independent tasks are studied in mobile edge computing systems with single user and multi core server.Firstly,the task offloading scheduling based on the hybrid flow shop scheduling model is modelled to calculate the system latency,based on it,the system energy consumption is modelled.Secondly,a hybrid encoding genetic algorithm is proposed to minimize the weighted sum of the system delay and energy consumption.Then,the optimal task offloading scheduling strategy and the trade off relationship between the system latency and the energy consumption are determined by the solution.Finally,simulation results provide the optimal Gantt chart of task offloading.It is shown that the system latency increases linearly with the number of tasks overall.Besides,compared with a random scheduling algorithm,the system latency decreases with the unloading strategy under the same conditions.In addition,a way to save energy without increasing system latency is found due to the inverse relationship between the system energy consumption and the system latency.
keywords:mobile edge computing  multi core server  task offloading scheduling  power control  hybrid flow shop scheduling  genetic algorithm
查看全文  查看/发表评论  下载PDF阅读器

你是第3000772访问者
版权所有《南京邮电大学学报(自然科学版)》编辑部
Tel:86-25-85866913 E-mail:xb@njupt.edu.cn
技术支持:本系统由北京勤云科技发展有限公司设计