Search research articles

关于 JoVE

概览领导团队博客 JoVE 帮助中心

作者

出版流程编辑委员会范围与政策同行评审常见问题投稿

图书馆员

用户评价订阅访问资源图书馆顾问委员会常见问题

研究

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments 存档

教育

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual 教师资源中心教师网站

使用条款与条件

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序

Same author

Ecophysiological Characteristics of High-Photosynthetic-Efficiency Rice Varieties and Their Environmental Regulation.

Physiologia plantarum·2026

Same author

Biomedical Knowledge Graph Alignment with GPT-Augmented Similarity Feature Construction via Tree-based Particle Swarm Optimization and Adaptive Fitness Optimization.

IEEE journal of biomedical and health informatics·2026

Same author

The ubiquitin-like protein FAT10 enhances the autophagy-mediated degradation of ZO-1 by stabilizing ATG3 to promote the lung metastasis of colon cancer.

Cellular oncology (Dordrecht, Netherlands)·2026

Same author

Multifunctional graphene oxide and PEDOTPSS nanocomposite enabling highly efficient blue multishelled ZnSeTe quantum dot light emitting diodes.

Scientific reports·2026

Same author

Quercetin Alleviates Renal Ischemia-Reperfusion Injury by Inhibiting the STING-NF-κB Inflammatory Pathway.

Dose-response : a publication of International Hormesis Society·2026

Same author

Preliminary observations on the effect of Kegel exercise and perihip muscle exercise training and combined training on transversus abdominis thickness in healthy young women.

Medicine·2026

Search research articles

相关实验视频

Updated: Jul 11, 2025

Large Scale Energy Efficient Sensor Network Routing Using a Quantum Processor Unit

Large Scale Energy Efficient Sensor Network Routing Using a Quantum Processor Unit

Published on: September 8, 2023

基于SDWN中的多代理深度强化学习的智能多播路由方法.

Hongwen Hu¹, Miao Ye^2,3, Chenwei Zhao²

¹School of Computer Science and Information Security, Guilin University of Electronic Technology, Guilin 541004, China.

Mathematical biosciences and engineering : MBE

|November 3, 2023

概括

本研究介绍了一种使用软件定义无线网络 (SDWN) 的多代理深度强化学习 (MADRL-MR) 的新型多播路由方法. 通过智能地适应网络变化,MADRL-MR可提高吞吐量并减少延迟.

关键词:

深度强化学习的学习.多个代理的多个代理.多播多播的多播.软件定义的无线网络.

更多相关视频

Automated Deployment of an Internet Protocol Telephony Service on Unmanned Aerial Vehicles Using Network Functions Virtualization

Automated Deployment of an Internet Protocol Telephony Service on Unmanned Aerial Vehicles Using Network Functions Virtualization

Published on: November 26, 2019

A Networked Desktop Virtual Reality Setup for Decision Science and Navigation Experiments with Multiple Participants

A Networked Desktop Virtual Reality Setup for Decision Science and Navigation Experiments with Multiple Participants

Published on: August 26, 2018

相关实验视频

Last Updated: Jul 11, 2025

Large Scale Energy Efficient Sensor Network Routing Using a Quantum Processor Unit

Large Scale Energy Efficient Sensor Network Routing Using a Quantum Processor Unit

Published on: September 8, 2023

Automated Deployment of an Internet Protocol Telephony Service on Unmanned Aerial Vehicles Using Network Functions Virtualization

Automated Deployment of an Internet Protocol Telephony Service on Unmanned Aerial Vehicles Using Network Functions Virtualization

Published on: November 26, 2019

A Networked Desktop Virtual Reality Setup for Decision Science and Navigation Experiments with Multiple Participants

A Networked Desktop Virtual Reality Setup for Decision Science and Navigation Experiments with Multiple Participants

Published on: August 26, 2018

科学领域:

计算机科学计算机科学
网络化网络化网络化
人工智能的人工智能

背景情况:

传统的无线网络在多播路由方面难以处理动态状态信息,从而影响服务质量 (QoS).
无线环境中的设备密度高加剧了高效多播通信的挑战.

研究的目的:

为软件定义无线网络 (SDWN) 环境提出一个新的多播路由方法,MADRL-MR.
为了提高吞吐量,减少延迟,并在密集的无线网络中提高多播路由效率.

主要方法:

利用软件定义无线网络 (SDWN) 实现灵活的网络配置和通过流量矩阵获取全球状态信息.
采用多代理深度增强学习 (MADRL) 来将多播路由划分为合作子问题.
基于流量,多播树状态和AP节点设计代理状态和动作空间,采用新的单跳动策略和四态奖励函数.

主要成果:

在吞吐量,延迟和数据包丢失率方面,MADRL-MR在现有算法上表现出优越的性能.
该方法在动态网络条件下成功建立了更智能的多播路线.
分散培训与转移学习加快了融合,提高了代理人的适应能力.

结论:

在密集,动态的无线网络中,MADRL-MR为多播路由提供了强大而智能化的解决方案.
SDWN和MADRL的集成为网络性能和QoS提供了显著的优势.
拟议的方法有效地解决了传统的多播路由方法的局限性.