Abstract: Mobile Edge Computing (MEC) is helpful for reducing computational burden and energy consumption during workflow mobilizaiton.However,more time and energy wastes could be caused by unreasonable task offloading.For this reason,a two-stage Mobile Edge Computing Offloading Optimization algorithm based on Dynamic Critical Path (MECOODCP) algorithm was introduced,which included mobile edge computing offloading optimization algorithm and dynamic critical path algorithm base on local computation.On this basis,the optimized edge unloading strategy was proposed,which  had taken advantage of the selected optimized eNB from Hidden Markov prediction with velocity and offset prediction to ensure the success rate of task offloading.Meanwhile,the critical path of workflow was dynamically updated to reduce the impact on scheduling results.The effectiveness of the proposed algorithm was proved by the simulation experiment,and the proposed algorithm could optimize the workflow under the mobile edge environment by decreasing 12% of time and 6% of energy compared with traditional optimization algorithm.

Key words: mobile edge, offloading optimization, optimized eNB, dynamic critical path, energy consumption, workflow

摘要： 移动边缘计算有助于减少工作流调动中用户终端的能耗和计算负担,但不合理的任务卸载会导致设备产生大量时间和能源的消耗。针对该问题,提出一种面向边缘侧卸载优化的工作流动态关键路径调度的两阶段算法,包括边缘侧卸载优化算法和基于本地计算量的动态关键路径调度算法。制定了边缘侧卸载优化的策略,该策略通过隐性马尔科夫预测得到可卸载eNB集并结合速度与偏移量预测筛选最优可调度eNB,以确保卸载成功率;同时在调度过程中通过动态更新关键路径,避免了关键路径变化对调度结果的影响。通过仿真实验证明了所提算法的有效性。相比传统优化算法,该算法能优化移动边缘环境下工作流12%的完工时间,并减少6%的能耗。

关键词: 移动边缘, 卸载优化, 最优可调度eNB, 动态关键路径, 能耗, 工作流