面向移动感知的计算卸载及资源分配策略研究  

Research on Mobility-Aware Computation Offloading and Resource Allocation Strategy

作　　者：班玉琦 段利国 温昊宇 李爱萍 赵菊敏 BAN Yuqi;DUAN Liguo;WEN Haoyu;LI Aiping;ZHAO Jumin(College of Information and Computer,Taiyuan University of Technology,Jinzhong 030600,Shanxi,China;College of Artificial Intelligence,Nanjing Agricultural University,Nanjing 210031,China)

机构地区：[1]太原理工大学信息与计算机学院,山西晋中030600 [2]南京农业大学人工智能学院,南京210031

出　　处：《计算机工程》2023年第8期163-173,共11页Computer Engineering

基　　金：国家自然科学基金(61972273)。

摘　　要：在移动边缘计算(MEC)中,用户设备将计算密集任务卸载至边缘服务器执行以降低执行时延与能耗,基于5G技术的新型应用要求计算过程支持设备的高速移动性,而目前计算卸载方案的研究大多集中于静态场景。为提高用户体验质量,在多设备与多MEC服务器场景下,对MEC中考虑设备移动轨迹的计算卸载方案进行研究。结合设备移动性、计算与通信资源、信道状态及任务需求等因素,将场景下的计算卸载方案设计为混合整数非线性规划问题。为降低求解难度,将上述问题分解为卸载服务器选择问题和固定服务器选择方案下的计算资源分配与子信道选择问题,采用凸优化技术及改进的Kuhn-Munkres算法对子问题进行求解,并依据子问题的解设计启发式卸载服务器选择算法,基于多项式时间复杂度获得次优卸载方案。通过EdgeCloudSim工具对本文卸载策略进行仿真,并与常用的卸载算法进行对比,实验结果表明,该算法在满足任务的实时性要求下,与穷举算法的平均系统效用差距控制在2.3%以内。In Mobile Edge Computing(MEC),user equipment offloads computationally intensive tasks to edge servers for execution to reduce execution delay and energy consumption.This process requires 5G technology-based applications to support the high-speed movement of devices during computing.However,much of the current research on computational offload solutions is focused on static scenarios.To improve the quality of user experience,this study investigates a computational offloading scheme that considers device movement trajectories in MEC and thus more suitable multi-device and multi-MEC server scenarios.Because this scheme considers multiple factors such as device mobility,computing and communication resources,channel states,and mission requirements,it can be described as a mixed-integer nonlinear programming problem.To reduce the difficulties inherent in solving this problem,this study decomposes the problem into subproblems of offloading server selection,computing resource allocation,and subchannel selection under a fixed-server selection scheme.The convex optimization technique and improved Kuhn-Munkres algorithm are then used to solve the subproblems.This study also designs a heuristic offload server selection algorithm based on the solution to the subproblems and derives a suboptimal offload solution with polynomial time complexity.Simulations are conducted using the EdgeCloudSim tool,the results of which prove the effectiveness of the proposed algorithm as compared with five other commonly used offloading algorithms.The experimental results show that the average system utility gap between the algorithm and exhaustive algorithm can be controlled to within 2.3%when it meets the real-time requirements of a given task.

关 键 词：移动边缘计算 移动感知 计算卸载 资源分配 卸载算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]