多云协同辅助车辆计算的资源管理策略  

Resource management strategy for multi-cloud collaboratively assisting vehicle computing

作　　者：张文萍 陈桂芬[1] ZHANG Wen-ping;CHEN Gui-fen(School of Electronics and Information Engineering,Changchun University of Science and Technology,Changchun 130022,China)

机构地区：[1]长春理工大学电子信息工程学院,吉林长春130022

出　　处：《计算机工程与设计》2021年第8期2180-2185,共6页Computer Engineering and Design

基　　金：吉林省科技厅基金项目(20190302103G X);武器装备预先研究基金项目(6141B021322、6141B012826)。

摘　　要：针对高质量车载应用海量增长的问题,结合车辆自身资源受限以及传统云计算资源使用成本高和时延高的问题,定义一种多云协同辅助车辆计算(multi-cloud collaboratively assisting vehicle computing,MCAVC)范式,提出基于对编码、交叉和变异操作做出改进的遗传算法(genetic algorithm,GA)的节点选择和资源分配联合优化方案,目标是使任务完成时间和使用计算资源的货币成本降到最低。实验结果表明,所提方案在时间和资源成本的加权和方面优于现有方案,改进算法相比实数编码GA和传统GA性能更优。Aiming at the problem of the massive growth of high-quality vehicle applications,combined with the limitation of vehicles’own resources and the high using cost and high delay of traditional cloud computing resources,a multi-cloud collaboratively assisting vehicle computing(MCAVC)paradigm was defined,and a joint optimization scheme of node selection and resource allocation based on genetic algorithm(GA)with improved coding,crossover and mutation operations was proposed.Its goal was to minimize the total computing overhead in terms of the weighted sum of task completion time and the monetary cost of using computing resources.The results show that the proposed scheme is superior to the existing scheme in terms of total computational cost,and the improved algorithm has better performance than real number coding GA and traditional GA.

关 键 词：车载应用 节点选择 资源分配 云计算 遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]