SDN中基于强化学习的多业务智能QoS路由方法  

Multi Service Intelligent QoS Routing Method Based on Reinforcement Learning in SDN

作　　者：翟凡妮 ZHAI Fanni(Department of Electronic Information,Shaanxi Institute of Technology,Xi'an 710300,China)

机构地区：[1]陕西国防工业职业技术学院电子信息学院,西安710300

出　　处：《智能物联技术》2025年第1期44-47,共4页Technology of Io T& AI

摘　　要：为了解决现有路由算法无法学习历史路由决策经验导致的网络负载不均衡问题,将强化学习技术引入软件定义网络(Software Defined Network,SDN)的服务质量(Quality of Service,QoS)路由问题,提出一种基于强化学习的多业务智能QoS路由方法MDQN(Multi-service QoS routing method based on DeepQ Network)。该方法部署在SDN控制器中,能学习历史决策经验,并在网络状态发生变化时及时调整路径。通过在SDN中部署该方法,有效平衡了网络负载,增加了网络的吞吐量,为SDN中的QoS路由问题提供了一种有效的解决方案。To address the issue of network load imbalance caused by existing routing algorithms'inability to learn from historical routing decision experiences,reinforcement learning technology is introduced into the Quality of Service(QoS)routing problem in Software Defined Network(SDN),and a multi-service intelligent QoS routing method based on reinforcement learning,Multi-service QoS routing method based on Deep Q Network(MDQN),is proposed.This method is deployed in the SDN controller,capable of learning from historical decision experiences and adjusting the path in a timely manner when network conditions change.By deploying this method in SDN,it effectively balances network load and improves network throughput,providing an effective solution to the QoS routing problem in SDN.

关 键 词：软件定义网络(SDN) 服务质量(QoS) 强化学习 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]