Efficient task completion for parallel offloading in vehicular fog computing
Efficient task completion for parallel offloading in vehicular fog computing
复制标题
DOI：
10.23919/jcc.2019.11.004
复制
发表时间：
2019-11
影响因子：
4.1
通讯作者：
Jindou Xie;Yunjian Jia;Zhengchuan Chen;Zhaojun Nan;Liang Liang-Liang
中科院分区：
文献类型：
--
作者：
Jindou Xie;Yunjian Jia;Zhengchuan Chen;Zhaojun Nan;Liang Liang-Liang
本文研究了车载雾计算系统，提出了一种有效的并行卸载方案。服务时间，解决了任务卸载延迟，任务分解和切换成本，作为卸载性能的度量。提出了一种基于可用资源感知的并行计算卸载方案，该方案综合考虑了车辆移动性和时变计算能力的影响，通过RSU确定目标节点进行计算卸载。该方案基于隐马尔可夫模型和马尔可夫链理论，通过联合实现移动感知和计算感知，有效地处理了雾节点选择过程中系统状态信息的不完备性。仿真结果证实了理论分析和算法的有效性。
In this paper, we investigate vehicular fog computing system and develop an effective parallel offloading scheme. The service time, that addresses task offloading delay, task decomposition and handover cost, is adopted as the metric of offloading performance. We propose an available resource-aware based parallel offloading scheme, which decides target fog nodes by RSU for computation offloading jointly considering effect of vehicles mobility and time-varying computation capability. Based on Hidden Markov model and Markov chain theories, proposed scheme effectively handles the imperfect system state information for fog nodes selection by jointly achieving mobility awareness and computation perception. Simulation results are presented to corroborate the theoretical analysis and validate the effectiveness of the proposed algorithm.