This website requires JavaScript.

Structure-Enhanced DRL for Optimal Transmission Scheduling

Jiazheng ChenWanchun LiuDaniel E. QuevedoSaeed R. KhosraviradYonghui LiBranka Vucetic
Dec 2022
摘要
Remote state estimation of large-scale distributed dynamic processes plays animportant role in Industry 4.0 applications. In this paper, we focus on thetransmission scheduling problem of a remote estimation system. First, we derivesome structural properties of the optimal sensor scheduling policy over fadingchannels. Then, building on these theoretical guidelines, we develop astructure-enhanced deep reinforcement learning (DRL) framework for optimalscheduling of the system to achieve the minimum overall estimation mean-squareerror (MSE). In particular, we propose a structure-enhanced action selectionmethod, which tends to select actions that obey the policy structure. Thisexplores the action space more effectively and enhances the learning efficiencyof DRL agents. Furthermore, we introduce a structure-enhanced loss function toadd penalties to actions that do not follow the policy structure. The new lossfunction guides the DRL to converge to the optimal policy structure quickly.Our numerical experiments illustrate that the proposed structure-enhanced DRLalgorithms can save the training time by 50% and reduce the remote estimationMSE by 10% to 25% when compared to benchmark DRL algorithms. In addition, weshow that the derived structural properties exist in a wide range of dynamicscheduling problems that go beyond remote state estimation.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答