基于边缘计算的基站端计算资源调度策略研究

Research on resource scheduling strategy of base station terminal computing based on edge computing

传统的移动边缘计算是将计算任务卸载到移动边缘服务器,特殊场景下如边缘网络带宽受限、基站覆盖能力变弱、甚至灾害或战争情况下导致基站与边缘服务器通信中断,复杂计算任务无法提交到边缘甚至云端,基站上行传输强行被切断。结合移动自组织网络以及雾计算技术,给出基站端计算技术,基站通过下行传输将计算任务转发给小区内的移动智能终端,利用资源调度策略实现完成任务计算。资源调度策略中,针对基站端计算网络下的终端管理,提出了一种通过衡量并发传输时延的可获得最大计算力的聚簇算法,利用终端可用计算力与簇总计算力的关系,比例分配计算任务和带宽,最大效能地提高频谱利用率,从而快速完成原有计算任务。仿真结果表明,本文算法在基站端计算网络中能有效减少系统时延,高效完成计算任务。

The traditional mobile edge computing offloads computing tasks to mobile edge servers.In special scenarios such as limited edge network bandwidth,weakened base station coverage,and even disaster or war situations,communication between base stations and edge servers is interrupted.Complex computing tasks cannot be submitted to the edge or even the cloud,and the uplink transmission of base stations is forcibly cut off.Combining mobile ad hoc networks and fog computing technology,a base station terminal computing technology is proposed.The base station forwards computing tasks to mobile intelligent terminals in the community through downlink transmission,and uses resource scheduling strategies to complete task calculations.In the resource scheduling strategy,a clustering algorithm is proposed for terminal management in the base station computing network,which can obtain the maximum computation by measuring the concurrent transmission delay.By utilizing the relationship between the available computation of the terminal and the total computation of the cluster,computing tasks and bandwidth are proportionally allocated,maximizing the efficiency of spectrum utilization and quickly completing the original computing tasks.The simulation results show that the algorithm proposed in this paper can effectively reduce system latency and efficiently complete computational tasks in the base station computing network.