设备间任务依赖的最佳卸载决策和资源分配

Optimal Offloading Decision and Resource Allocation of Inter-devices Task Dependency

对于不同设备之间具有任务依赖性的问题,考虑了两个设备的移动边缘计算(Mobile Edge Computing,MEC)与端对端(Device-to-Device,D2D)技术协作网络,其中一个无线设备的最终输出作为另一个设备上某个任务的输入。在此任务依赖模型下,为了最小化无线设备的能耗和任务完成时间的加权和,研究了最佳的资源分配(卸载发射功率和本地CPU频率)和任务卸载决策问题。为了解决该问题,将原问题分解为给定任务卸载决策的资源分配问题和优化与资源分配问题相对应的任务卸载问题。首先给定卸载决策,推导出卸载发射功率和本地CPU频率的闭合表达式,运用凸优化方法求出该问题的解。然后证明最优卸载决策遵循一次爬升策略,在此基础上提出了一种降低复杂度的在线任务卸载算法,该算法可以在多项式时间内获得最优卸载决策。数值结果表明,该策略的性能明显优于其他有代表性的基准测试,同时MEC与D2D协作可以显着提高系统的性能。

For the problem of task dependence between different devices,the mobile edge computing(MEC)and device-to-device(D2 D)technology cooperation network of two devices are considered.The final output of one wireless device is the input of a task on another device.In this task-dependent model,in order to minimize the weighted sum of energy consumption of wireless devices and task completion time,the optimal resource allocation(offloading transmitting power and local CPU frequency)and task offloading decision-making are studied.In order to solve this problem,the original problem is decomposed into a resource allocation problem for a given task offloading decision and a task offloading problem corresponding to the optimization and resource allocation problem.Firstly,given the offloading decision,the closed-form expressions of the offloading transmission power and the local CPU frequency are derived,and the convex optimization method is used to solve the problem.Then it is proved that the optimal offloading decision follows a climbing strategy,and on this basis,an online task offloading algorithm with reduced complexity is proposed,which can obtain the optimal offloading decision in polynomial time.The numerical results show that the performance of this strategy is significantly better than that of other representative benchmark tests,and the collaboration between MEC and D2 D can significantly improve the performance of the system.