结合无线能量传输的移动边缘计算系统高能效资源配置

Energy Efficient Resource Allocation for Wireless Power Transfer Powered in Mobile Edge Computing Systems

利用无线能量传输(Wireless Power Transfer,WPT)向移动边缘计算(Mobile Edge Computing,MEC)系统中的用户设备供电以进一步提高系统的能效和计算可持续性是未来潜在的研究方向。提出了一种新的基于双阶段的深度Q网络(Deep Q-Network,DQN)优化框架。该框架求解了长期优化问题的同时也减小了强化学习的动作空间,提升了策略性能。通过大量仿真验证,提出的双阶段DQN方案具有更快的收敛速度,并且能实现接近60%的MEC系统能耗降低。

Introducing the base station supporting Wireless Power Transfer(WPT) into Mobile Edge Computing(MEC) system to further improve the energy efficiency and computing sustainability of user equipment is a potential research direction in the future.This paper proposes a new two-stage optimization scheme.The long-term optimization problem is solved,the action space of reinforcement learning is reduced, and the strategy performance is improved.Through extensive simulations,the proposed two-stage DQN is shown to have a higher energy efficiency and faster convergence speed,and can achieve nearly 60% gain of energy efficiency.