基于遗传算法的MEC任务卸载资源优化策略  

Resource optimization strategy of MEC task unloading based on genetic algorithm

作　　者：王丽[1] 王晓凯[2] WANG Li;WANG Xiao-kai(Department of Physical and Electronic Engineering,Jinzhong University,Jinzhong 030619,China;School of Physics and Electronic Engineering,Shanxi University,Taiyuan 030006,China)

机构地区：[1]晋中学院物理与电子工程系,山西晋中030619 [2]山西大学物理电子工程学院,山西太原030006

出　　处：《计算机工程与设计》2023年第10期2909-2915,共7页Computer Engineering and Design

基　　金：山西省重点研发基金项目(高新技术领域)(201803D121102);山西省高等学校教学改革创新基金项目(J2021646)。

摘　　要：为解放计算能力和能量资源有限的移动终端设,引入边缘计算加快物联网真实场景中终端数据的处理和分析过程。在网络边缘处理工作负载可以减少移动边缘计算的延迟,但会大大增加系统功耗,因此迫切需要改进网络边缘服务器的能量模型。对该模型进行改进,减少安全物联网系统中多传感器框架的延迟和功耗问题。使用遗传算法(genetic algorithm,GA)处理大量请求以及相应的延迟和功耗限制。仿真结果表明,当传输到终端节点的工作负载为2 MB时,几种方法总延迟将保持在2 s左右,所提算法在不同工作负载下的功耗和延迟方面都有更显著的降低。To liberate the mobile terminal equipment with limited computing power and energy resources,edge computing was introduced to speed up the processing and analysis of terminal data in the real scene of the internet of things.Processing workload at the network edge can reduce the latency of mobile edge computing,but it will greatly increase the system power consumption.Therefore,it is urgent to improve the energy model of network edge server.The model was improved to reduce the delay and power consumption of multi-sensor framework in secure internet of things system.Genetic algorithm(GA)was used to deal with a large number of requests and the corresponding quality and security constraints.The simulation results show that when the workload transmitted to the terminal node is 2 MB,the total delay of several methods is maintained at about 2 s,and the proposed algorithm has more significant reduction in power consumption and delay under different workloads.

关 键 词：物联网 移动边缘计算 遗传算法 计算资源 终端数据 传感器 系统功耗 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]