基于非支配排序遗传策略的车联网多目标计算任务卸载调度方法  

An approach of multi-objective computing task offloading scheduling based on NSGS for IOV

作　　者：张德干[1] 张志昊 张捷 张婷 朴铭杰 姜星如 ZHANG Degan;ZHANG Zhihao;ZHANG Jie;ZHANG Ting;PIAO Mingjie;JIANG Xingru(Tianjin Key Lab of Intelligent Computing&Novel software Technology,Tianjin University of Technology,Tian-jin 300384,China;School of Electronic and Information Engineering,Beijing Jiaotong University,Beijing 100044,China;School of Sports Economics and Management,Tianjin University of Sport,Tianjin 301617,China;Col-lege of Software,Beihang University,Beijing 100083,China)

机构地区：[1]天津理工大学天津市智能计算及软件新技术重点实验室,天津300384 [2]北京交通大学电子信息工程学院,北京100044 [3]天津体育学院体育经济与管理学院,天津301617 [4]北京航空航天大学软件学院,北京100083

出　　处：《北京交通大学学报》2023年第2期45-57,共13页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：国家自然科学基金(61571328);天津市自然科学基金(18JCZDJC96800);天津市重大科技专项基金(17YFZCGX00360)。

摘　　要：移动边缘计算(Mobile Edge Computing,MEC)作为5G体系结构中非常重要的部分,能够支持需要超低延迟的许多创新性的服务与应用,可以通过引入MEC来解决目前车载移动终端设备无法满足车联网(Internet of Vehicles,IoV)低能耗与低时延需求的问题.提出将车辆计算任务切分成小的有依赖关系的子任务,切分后的子任务可并行处理,同时基于计算任务切分提出时延与能耗模型;构建IoV计算任务卸载的约束多目标优化模型,并提出非支配排序遗传策略(Nondominated Sorting Genetic Strategy,NSGS)来优化目标函数,对IoV中计算任务卸载问题提出新的非支配关系与约束.基于一系列的实验以及卸载方法间的比较,证明了本文所提出方法的有效性及具有更低的时延和能耗.Mobile Edge Computing(MEC),as a very important part of the 5G architecture,can support many innovative services and applications that require ultra-low latency,and the introduction of MEC can solve the problem of low energy consumption and low latency for the Internet of Vehicles(IoV),which cannot be met by current in-vehicle mobile devices.In this paper,the vehicle computing task is proposed to divided into two parts.Then the vehicle computing task is diveded into small dependent subtasks,which can be processed in parallel,and a latency and energy consumption model is proposed based on this task division;a constrained multi-objective optimisation model is constructed for IoV computational task unloading,a constrained multi-objective optimization model is also proposed for computational task offloading in IoV,and a Nondominated Sorting Genetic Strategy(NSGS) is adopted to optimize the objective function,and new nondominated relations and constraints for the computational task offloading problem in IoV.Based on a series of experiments and comparisons between offloading methods,the effectiveness of the proposed method and its lower latency and energy consumption are demonstrated.

关 键 词：车联网 移动边缘计算 计算任务卸载 任务切分 非支配排序遗传策略 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]