无线供能协作计算网络中的时延最小化资源分配方案

Latency minimization resource allocation scheme in wireless powered cooperative computing network

在单专用能量站为多个用户无线供能的场景下,为缩短任务处理时延,设计了一种新型多用户协作计算方案。建立了关于匹配决策和资源分配的优化问题,在用户间一对一匹配情况下提出了一种基于交替优化和匈牙利算法的高性能求解方案和一种基于重构线性化方法的低复杂度求解方案;针对一对多匹配情况提出了一种改进的贪婪算法。实验结果表明,在用户间一对一匹配时所提方案能够较对比方案降低最多12.6%的任务处理时延;一对多匹配情况下所提方案节省了5%的任务处理时延,即所提方案能有效保障用户端的时延需求。

In the scenario that a dedicated energy transmitter serves multiple wireless-powered users,aiming to minimize the task execution latency,this paper proposed a new multi-user cooperative computation scheme and formulated the problems to optimize user matching and resource allocation.For the one-to-one matching case between users,this paper derived a high-performance solution based on the alternating optimization and Hungarian algorithms and a low-complexity solution based on the reformulation-linearization technique.For the one-to-many matching case,this paper derived an improved greedy algorithm.Experimental results show that in the case of one-to-one matching,the proposed scheme can reduce the task processing delay by up to 12.6%compared with the comparison schemes,for the one-to-many matching case,the proposed scheme saves 5%of the task processing delay.The proposed scheme can meet the task execution latency requirements of the users.