| 摘要: |
| 针对在任务卸载时由于设备的移动而导致任务迁移这一问题，将任务卸载过程建模为马尔科夫决策过程，并通过优化资源分配和任务卸载策略，解决基于联合时延和能耗的损耗函数最小的优化问题。首先将问题转化为最小化损耗函数之和，并在决策前对每个任务的传输功率采用二分法进行优化，然后基于获得的传输功率提出一种QLBA（Q-learning Based Algorithm）来完成卸载决策。仿真结果证实所提方案优于传统算法。 |
| 关键词: 移动边缘计算(MEC) 资源分配 任务迁移 马尔科夫决策过程 Q-learning |
| DOI： |
|
| 基金项目:国家重点研发计划跨政府合作专项（2016YFE0118400）；河南省自然科学基金（202300410482）；郑州市重大科技创新专项(2019CXZX0037) |
|
| Resource allocation and task offloading strategy in mobile edge computing system for task migration |
| JIA Shuxia,HAO Wanming,GAO Zihan,YANG Shouyi |
| (School of Information Engineering,Zhengzhou University,Zhengzhou 450001,China) |
| Abstract: |
| For the problem of task migration caused by the movement of equipment during task offloading,the task offloading process is modeled as a Markov decision process(MDP) and the optimization problem of Loss Function(LF) based on joint delay and energy consumption is solved by optimizing resource allocation and task offloading strategy.Firstly,the problem is transformed into the sum of the minimum LF and the transmitted power of each task is optimized by bisection method before the decision is made.Then a Q-learning Based Algorithm(QLBA) is proposed based on the acquired transmitted power to complete the offloading decision.The simulation results show that the proposed scheme is superior to the traditional algorithm. |
| Key words: mobile edge computing(MEC) resource allocation task migration Markov decision process(MDP) Q-learning |
