面向优先级用户的移动边缘计算任务调度策略

随着移动终端处理的数据量及计算规模不断增加，为降低任务处理时延、满足任务的优先级调度需求，结合任务优先级及时延约束，提出了基于任务优先级的改进min-min调度算法（task priority-based min-min，TPMM）。该算法根据任务的处理价值及任务的数据量计算任务的优先级，结合任务截止时间、服务器调度次数制定资源匹配方案，解决了边缘网络中服务器为不同优先级的用户进行计算资源分配的问题。仿真实验结果表明，该算法可以均衡服务器利用率，并有效降低计算处理的时延，提高服务器在任务截止处理时间内完成任务计算的成功率，相比min-min调度算法，TPMM算法最多可降低78.45%的时延，提高80%的计算成功率；相比max-min调度算法，TPMM算法最多可降低80.15%的时延并提高59.7%的计算成功率；相比高优先级（high priority first，HPF）调度算法，TPMM算法最多降低59.49%的时延，提高57.7%的计算成功率。

Task scheduling policy for mobile edge computing with user priority

With the increasing of data and the calculation scale in the mobile terminal, this paper proposed TPMM based on the min-min scheduling algorithms to reduce the task processing delay and meet the priority scheduling requirements of the task, it also combined the task priority and deadline constraints. The algorithm calculated the priority of the task according to the processing value and data volume, then formulated the resource matching scheme according to the task deadline and the number of server scheduling times, to solve the problem that the server in the edge network allocated resources for users with different priorities. Simulations show that the algorithm can balance the server utilization, effectively reduce the delay of the calculation process and improve the success rate of the server to complete the task calculation within the task deadline processing time. Comparing with the min-min scheduling algorithm, TPMM algorithm can reduce the delay by 78.45% and increase the calculation success rate by 80%. Comparing with the max-min scheduling algorithm, TPMM algorithm can reduce the delay by 80.15% and improve the calculation success rate by 59.7%. Comparing with HPF scheduling algorithm, TPMM algorithm reduces the delay up to 59.49% and increases the computational success rate by 57.7%.