摘要
为了解决现代骨干网络在处理大流量数据时面临的流量负载平衡问题,提出了一种基于深度强化学习(DRL)的自主决策优化技术,其通过智能化调整流量分配来优化网络性能。采用实时数据驱动的学习模型,通过分析历史与当前网络负载数据,使系统能够在无需预定模型的情况下,学习如何动态调整流量以适应网络状态的变化。该技术有效整合了网络状态的实时监测、流量预测及自适应路由算法,形成了一个能够预测并响应网络状况变化的多层次决策系统,对高流量网络环境下的流量管理提供了一种有效的解决方案。
To address the issue of traffic load balancing in modern backbone networks when handling large data volumes,it proposes an autonomous decision optimization technique based on Deep Reinforcement Learning(DRL),which optimizes network performance through intelligent traffic distribution adjustments.The research employs a real-time data-driven learning model that analyzes both historical and current network load data,enabling the system to learn how to dynamically adjust traffic without the need for a predefined model,in order to adapt to changes in network conditions.This technique effectively integrates real-time monitoring of network status,traffic prediction,and adaptive routing algorithms,forming a multi-level decision system capable of predicting and responding to changes in network conditions,which provides an effective solution for traffic management in high-traffic network environments.
作者
李海彬
林雨浓
Li Haibin;Lin Yunong(China Unicom Yunnan Branch,Kunming 650051,China)
出处
《邮电设计技术》
2024年第5期19-24,共6页
Designing Techniques of Posts and Telecommunications
关键词
骨干网络
流量负载平衡
深度强化学习
自主决策优化
实时
多层次决策系统
Backbone network
Traffic load balancing
Deep reinforcement learning
Autonomous decision optimization
Real-time
Multilevel decision system