基于强化学习的移动边缘计算任务卸载方法  

Offload Method of Mobile Edge Computing Task Based on Reinforcement Learning

作　　者：陈刚 王志坚 徐胜超 CHEN Gang;WANG Zhijian;XU Shengchao(School of Data Science,Guangzhou Huashang College,Guangzhou 511300,China)

机构地区：[1]广州华商学院数据科学学院,广州511300

出　　处：《计算机测量与控制》2023年第10期306-311,316,共7页Computer Measurement &Control

基　　金：国家自然科学基金面上项目(61772221);广州华商学院校内导师制科研项目资助(2023HSDS07)。

摘　　要：最佳卸载策略直接影响移动计算任务卸载的时延与能耗,因此提出基于强化学习方法的移动边缘计算任务卸载方法;首先对移动设备的计算任务卸载形式展开具体分析,并基于分析结果获取计算任务卸载能量消耗、发射功率、传输速率等相关参数值,以此建立移动边缘计算任务卸载模型;最后基于建立的卸载模型结合Q-Learning算法对计算任务实施强化学习,找出计算任务的最佳卸载策略,从而实现移动边缘计算任务的实时卸载;实验结果表明,使用强化学习方法开展移动边缘计算任务卸载时,卸载能耗低,仅为0.87 J、时延小,均不超过6 s。Optimal offloading strategy directly affects the latency and energy consumption of mobile computing task offloading.Therefore,a mobile edge computing task offloading method based on reinforcement learning is proposed.Firstly,the specific analysis of task offloading form for mobile devices is carried out,and based on the analysis results,the parameters such as energy consumption,transmission power,and transmission rate for task offloading are obtained,thereby establishing a mobile edge computing task offloading model.Finally,based on the established offloading model and combined with the Q-Learning algorithm,reinforcement learning is applied to determine the optimal offloading strategy for computing tasks,thus achieving the real-time offloading of mobile edge computing tasks.Experimental results show that when using reinforcement learning for the mobile edge computing task offloading,the offloading energy consumption is low,only 0.87 J,and the latency is small,no more than 6 s.

关 键 词：强化学习方法 Q-Learning算法 移动边缘 计算任务卸载 卸载模型 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]