多用户D2D计算卸载与资源分配算法

Multi-user D2D Computation Offloading and Resource Allocation Algorithm

基于边缘服务器架构的移动计算卸载场景避免了核心网链路的网络延迟,在一定程度上提高了计算任务的处理效率,然而在移动用户密集分布的场景下,该计算卸载架构可能面临较高的蜂窝网络接入延迟。为提高移动边缘计算的用户体验,本文在移动用户密集分布的场景下研究了基于D2D通信的多用户计算卸载问题。根据是否具有任务处理需求将用户分为两类,定义拥有密集型计算任务需要处理的用户为需求用户,无任务需要处理且能够提供计算资源的用户为空闲用户。需求用户可以将任务以整体的形式卸载至空闲用户进行辅助处理。联合考虑需求用户与空闲用户在计算卸载中的收益,分别构建了各方的效用函数并形成了相应的优化问题。为了提高各方用户的效用,本文提出了基于Stackelberg博弈与遗传算法相结合的计算卸载与资源分配算法,该算法为两层结构,内层基于Stackelberg博弈对需求用户计算资源租赁单价决策以及空闲用户计算资源分配决策进行优化,并证明了各方用户的策略存在唯一纳什均衡。外层采用遗传算法对需求用户的任务卸载决策进行优化,根据内层算法的反馈结果,可以求解出能够有效提高需求用户效用的任务卸载决策。为验证算法的性能,本文进行了大量的实验仿真工作,仿真结果表明,与其他方案相比,本文算法能够有效提高需求用户与空闲用户的效用。同时,本文还对不同参数对算法有效性的影响进行了仿真分析。

The mobile computation offloading scenario based on an edge server architecture avoids the network latency of core network links and improves the processing efficiency of computing tasks to a certain extent.However,in a scenario where mobile users are densely distributed,this computation offloading architecture may face a high cellular-network access delay.In order to improve the experience of mobile users,this study investigated the multi-user computation offloading problem based on D2D communication in a scenario where mobile users are densely distributed.Users were divided into two categories based on whether they had task processing needs.Users with intensive computing tasks to handle were defined as demand users,while users without tasks to handle and who could provide computing resources were defined as idle users.It was assumed that the task of a demand user could be offloaded to one idle user in a binary manner for auxiliary processing.Jointly considering the benefits to demand users and idle users,we constructed a utility function for each user and formulated corresponding optimization problems.A computation offloading and resource allocation algorithm was proposed based on the Stackelberg game and genetic algorithm to improve each user’s utility.The algorithm had a two-layer structure,with the inner layer optimizing the computing resource rental unit price decisions of demand users and computing resource allocation decisions of idle users based on the Stackelberg game.It could also be proved that a unique Nash equilibrium existed for the strategies of the demand and idle users.The outer layer used a genetic algorithm to optimize the task offloading decisions of the demand users.The optimal task offloading decisions of the demand users could be solved based on the feedback results of the inner layer.To verify the performance of the algorithm,an experimental simulation was conducted in this study.The simulation results showed that compared with other schemes,the algorithm proposed in this paper could effectively improve the utilities of demand and idle users.At the same time,in order to verify the effectiveness of our proposed algorithm,we also conducted a simulation analysis of the impacts of different parameters.