| 摘要: |
| 移动边缘计算(Mobile Edge Computing,MEC)通过将云计算能力下沉至用户侧，提高了用户的任务执行能力。但在热点小区中，MEC服务器存在计算资源有限的问题。为了减少热点小区内任务执行总代价，提出了一种基于主从MEC系统的任务联合卸载方案。首先，方案随机生成卸载集，然后将卸载集内的任务分配至目标MEC服务器执行。为此提出基于贪婪的多MEC选择算法(Greedy Based MultiMEC Selection Algorithm,GBMS)，并通过求解凸函数完成计算资源分配。最后，根据任务的本地计算代价与卸载计算代价更新卸载集，进一步降低总代价。仿真结果表明，联合卸载方案可缓解热点小区计算资源有限的问题，相比其他方案可以有效降低热点小区内任务执行总代价。 |
| 关键词: 移动边缘计算（MEC） 计算资源 卸载方案 资源分配 |
| DOI： |
|
| 基金项目:国家自然科学基金资助项目(61602073) |
|
| A master-slave MEC server collaborative offloading and resource allocation scheme |
| XIAN Yongju,SONG Qingyun,GUO Chenrong,LIU Chuang |
| (School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China) |
| Abstract: |
| Mobile Edge Computing(MEC) improves the users task execution ability by sinking cloud computing power to the user side.But in hotspot communities,MEC servers have the problem of limited computing resources.In order to reduce the total cost of task execution in hotspot communities,a joint task offloading scheme based on master-slave MEC system is proposed.Firstly,the scheme randomly generates the offloading set.Then,the tasks in the offloading set are assigned to the target MEC server for execution.To this end,a Greedy Based Multi-MEC Selection Algorithm(GBMS) is proposed.In this process,the calculation of resource allocation is completed by solving the convex function.Finally,the offloading set is updated according to the tasks local and offloading calculation cost,which can further reduce the total cost.Simulation results show that the joint offloading scheme can alleviate the problem of limited computing resources in hotspot communities,and also effectively reduce the total cost of task implementation in hotspot communities compared with other schemes. |
| Key words: mobile edge computing（MEC） computing resource offloading scheme resource allocation |
