基于深度强化学习的电力物联网动态切片策略研究

Dynamic Slicing Strategy for Power Internet of Things Based onDeep Reinforcement Learning

软件定义电力物联网支持构建承载不同业务的网络切片(Network Slice,NS),通过部署NS为具有业务需求的物联网设备提供端到端服务。业务NS的部署涉及2个互相耦合的问题,即虚拟网络功能(Virtual Network Function,VNF)部署和业务传输路由确定。在海量业务需求与动态网络场景中,NS部署方案需要根据网络状态,实现智能的动态灵活部署。针对上述问题,研究动态网络场景下的切片策略,基于深度强化学习算法求解VNF部署和业务传输路由确定这一复杂联合优化问题,实验证明所提策略能根据目前的网络状态灵活地改变部署方案,控制业务路由平均能量损耗、平均可靠性和平均剩余带宽占有率,提高了网络整体传输性能。

Software-defined power internet of things supports the construction of Network Slice(NS)that can carry different business requirements.By deploying NS,end-to-end services can be provided to internet of things devices with specific business demands.The deployment of business NS involves two interrelated issues,namely the deployment of Virtual Network Function(VNF)and the determination of business transmission routings.In dynamic network scenario with massive business requirements,NS deployment solutions need to achieve intelligent and dynamically flexible deployment based on the network status.To address the aforementioned problems,the slicing strategy in the dynamic network scenario is explored and the complex joint optimization problem of VNF deployment and business transmission routing determination is solved based on deep reinforcement learning algorithm.The experimental findings demonstrate that the proposed strategy effectively adjusts the deployment plan based on the current state,controls the average energy loss,average reliability,and average remaining bandwidth occupation of the service path,and improves the overall transmission performance of the network.