| 摘要: |
| 当前，多数任务卸载策略只考虑单边缘或者“物-边-云”的卸载方式，而没有对异地边缘服务器的资源进行充分利用。针对上述问题，提出了一种多边缘协作的网络架构，该架构中的任务可以选择在本地执行、本地服务器执行、异地服务器执行或者在云端执行。分别对4种执行方法的时延和能耗的加权求和建立数学模型。在传统的任务属性中引入新变量——终端所能承受的最大合作成本，以便吸引更多的异地边缘服务器积极协作完成终端任务的计算。针对传统的粒子群算法容易早熟和陷入局部最优的缺点，采用免疫粒子群优化算法(Immune Particle Optimization,IPSO)来对优化目标进行求解。仿真结果表明，与本地卸载策略、免疫算法(Immune Algorithm,IA)和粒子群(Particle Swarm Optimization,PSO)算法相比，所提任务卸载策略的总代价分别减少了66.7%，54%和45.5%，可以提高任务的执行效率，有效地减少系统的总代价。 |
| 关键词: 动边缘计算(MEC) 任务卸载 多边缘协作 免疫粒子群算法 合作成本 |
| DOI：10.20079/j.issn.1001-893x.220526003 |
|
| 基金项目:国家自然科学基金资助项目（61661018）;江苏省基础研究计划青年基金项目（BK20210064） |
|
| A Multi-edge Collaborative Task Offloading Strategy |
| XU Yongjie,LI Hui,,LAN Song,XU Wenxiao,YU Xinyuan |
| (1.College of Electronics and Information Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,China;2.College of Electronics and Information Engineering,Wuxi University,Wuxi 214015,China) |
| Abstract: |
| At present,most task offloading strategies only consider single-edge or object-edge-cloud offloading models and don’t make full use of the resources of remote edge servers.For the problem,a multi-edge collaborative network architecture is proposed.Tasks in this architecture can be executed locally,on local servers,on offsite servers,or in the cloud.A mathematical model is established for the weighted summation of delay and energy consumption of the four execution methods.In addition,in order to attract more remote edge servers to actively cooperate to complete the calculation of terminal task,a new variable,or the maximum cooperation cost that the terminal can bear,is added to the traditional task attribute.In view of the disadvantages of traditional particle swarm optimization(PSO) algorithm which is easy to premature and fall into local optimum,the authors adopt Immune Particle Swarm Optimization(IPSO) algorithm to accomplish the optimization goal.Simulation results show that,compared with local offloading strategy,Immune Algorithm(IA) and PSO，the total cost of IPSO is reduced by 66.7%,54% and 45.5% respectively.IPSO can improve the task execution efficiency and effectively reduce the total cost of the system. |
| Key words: mobile edge computing（MEC） task offloading multi-edge collaboration immune particle swarm optimization(IPSO) cost of cooperation |
