基于强化学习的IPv6校园网双栈流量调度方法

Dual-stack traffic scheduling method of IPv6 campus network based on reinforcement learning

由于传统方法在IPv6校园网双栈流量调度应用中效果不佳,网络吞吐率比较低,文章提出基于强化学习的IPv6校园网双栈流量调度方法,利用无向图建立IPv6校园网双栈链路拓扑模型,描述双栈链路负载均衡状态,以负载均衡度最大化、双栈链路路径长度最小化为目标建立目标函数,利用强化学习网络模型对目标函数求解,将最优调度策略反馈到模型中执行,以此实现基于强化学习的IPv6校园网双栈流量调度。实验证明,在设计方法应用下IPv6校园网吞吐量达到141.02 Gbps,在IPv6校园网双栈流量调度具有良好的应用前景。

Because the traditional method is not effective in dual-stack traffic scheduling application of IPv6 campus network and the network throughput is relatively low,a dual-stack traffic scheduling method of IPv6 campus network based on reinforcement learning is proposed.The dual-stack link topology model of IPv6 campus network was established by using undirected graph to describe the load balancing state of dual-stack link.The objective function was established by maximizing the load balancing degree and minimizing the path length of dual-stack link.The objective function was solved by using the reinforcement learning network model,and the optimal scheduling strategy was fed back into the model for execution.In this way,the dual stack traffic scheduling of IPv6 campus network based on reinforcement learning is realized.The experimental results show that the throughput of IPv6 campus network reaches 141.02 Gbps under the design method,which has a good application prospect in dual-stack traffic scheduling of IPv6 campus network.