基于多Agent强化学习的电力通信网跨层保护方法  

Cross Layer Protection Method for Power Communication Network Based on Multi Agent Reinforcement Learning

作　　者：陈毅龙 CHEN Yi-long(State Grid Shanghai Municipal Electric Power Company,Shanghai 200000 China)

机构地区：[1]国网上海市电力公司,上海200000

出　　处：《自动化技术与应用》2024年第10期112-115,共4页Techniques of Automation and Applications

基　　金：上海市重点实验室专项基金项目(05DZ33205)。

摘　　要：针对当前方法存在数据传输成功率低、传输延迟时间长以及开销大等题,设计基于多Agent强化学习的电力通信网跨层保护方法。首先使用多Agent强化学习算法设定网络多路径协议,控制网络节点数据接收能力,然后构建网络跨层安全构架,设定相应网络模型作为网络跨层保护的基础,最后使用罚函数法对模型进行求解,保证函数解具有较高的可靠性,根据求解结果实现对网络跨层算法的优化,实现电力通信网跨层保护方法。实验结果可知,所提方法的收包率得到了明显提升,传输延迟时间缩短,开销低。Aiming at the problems of low success rate of data transmission,long transmission delay and high overhead in current methods,a cross layer protection method for power communication network based on multi-agent reinforcement learning is designed.First,it uses the multi-agent reinforcement learning algorithm to set the network multipath protocol and control the data receiving capaci-ty of network nodes,then builds the network cross layer security architecture,sets the corresponding network model as the basis of the network cross layer protection,and finally uses the penalty function method to solve the model to ensure that the function solution has high reliability.Based on the solution results,realizes the optimization of the network cross layer algorithm and real-izes the cross layer protection method of power communication network.The experimental results show that the packet receiving rate of the proposed method is significantly improved,the transmission delay is shortened,and the overhead is low.

关 键 词：多AGENT强化学习 跨层保护 罚函数 数据包传输延迟 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TN915[自动化与计算机技术—控制科学与工程]