面向缓存机制的移动边缘计算任务卸载研究  

Research on Task Offloading of Mobile Edge Computing for Caching Mechanism

作　　者：陈姣 沈艳 CHEN Jiao;SHEN Yan(School of Computer Science,Chengdu University of Information Technology,Chengdu 610200,Sichuan,China)

机构地区：[1]成都信息工程大学计算机学院,四川成都610200

出　　处：《计算机工程》2024年第7期194-203,共10页Computer Engineering

基　　金：国家自然科学基金(62172061)。

摘　　要：在移动边缘计算(MEC)环境下,用户需求快速增长,但由于移动设备的计算和存储资源受限,时延和能耗问题日益突显。此外,任务的重复卸载和处理也进一步加剧了时延和移动设备的高能耗问题。针对上述问题,提出一种带有缓存机制的移动边缘计算任务卸载方案来减少任务卸载过程中的时延和能耗。首先,基于任务的流行度、新鲜度、数据大小等因素设计一个缓存机制,根据该机制的结果设计缓存更新策略。然后,针对任务卸载和缓存问题,建立一个联合优化模型,优化模型以最小化系统总成本为目标,考虑了任务卸载和缓存对移动设备时延和能耗的影响。为求解这一复杂的优化模型,通过添加惩罚函数的形式将约束条件加入目标函数,并采用粒子群优化(PSO)算法获得任务卸载和缓存的最优决策。实验结果表明,与传统的无缓存的本地计算、无缓存的任务卸载、使用随机缓存的任务卸载等方案相比,该方案的总时延降低了37.00%以上,缓存命中率提高了7.78%以上,具有较高的缓存资源利用率。As the Mobile Edge Computing(MEC)environment faces increasing user demands,the problems of delay and energy consumption stemming from the limited computing and storage resources of mobile devices are becoming increasingly prominent.These issues are further exacerbated by the repeated offloading and processing of tasks in mobile devices.To minimize the delay and energy consumption of task offloading,an MEC task offloading scheme with a caching mechanism is proposed herein.Firstly,a cache content selection model is established based on factors such as task popularity,freshness,and data size,and cache update strategies are formulated according to the results obtained.Then,a joint optimization model is established to address the issues of task offloading and caching,aiming to minimize the total system cost by considering the impact of task offloading and caching on device delay and energy consumption.To solve this complex optimization problem,the objective function is constrained by penalty functions and solved using the Particle Swarm Optimization(PSO)algorithm.The experimental results show that compared with traditional methods such as LOCal computing without caching(LOC),task OFFloading without caching(OFF),and task offloading using random caching,the total delay of this scheme is reduced by more than 37.00% while the cache hit rate is increased by more than 7.78%,indicating a high utilization rate of cache resources.

关 键 词：移动边缘计算 任务卸载 缓存机制 缓存更新 缓存命中率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]