基于Q-learning改进蜘蛛猴算法的电工装备边缘网关部署研究  

Research of the Deployment of Electrical Equipment Edge Gateway Based on Improved Spider Monkey Algorithm Using Q-Learning Reinforcement Learning

作　　者：吕振辉 张敬伟 崔强 陈曦 王胜 孙毅[4] 陈明昊 LV Zhenhui;ZHANG Jingwei;CUI Qiang;CHEN Xi;WANG Sheng;SUN Yi;CHEN Minghao(State Grid Shanghai Municipal Electric Power Company,Shanghai 200122,China;State Grid Materials Co.,Ltd.,Beijing 100120,China;Beijng Guo Dian Tong Network Technology Co.,Ltd.,State Grid Info&Telecom Group Co.,Ltd.,Beijing 100052,China;School of Electrical and Electronic Engineering,North China Electric Power University,Beijing 102206,China)

机构地区：[1]国网上海市电力公司,上海200122 [2]国网物资有限公司,北京100120 [3]国网信通产业集团北京国电通网络有限公司,北京100052 [4]华北电力大学电气与电子工程学院,北京102206

出　　处：《电力信息与通信技术》2022年第1期51-60,共10页Electric Power Information and Communication Technology

基　　金：国家电网有限公司总部科技项目资助“面向工业互联网的电工装备智慧物联体系研究与应用”(52090020002W)。

摘　　要：电工装备智慧物联平台通过部署在电力通信主干网路由器上的边缘网关对供应商数据汇集节点进行数据抓取与上传,实现平台与供应商的数据高效交互。面向海量生产数据和多个供应商数据汇集节点,文章以最小化数据抓取与转发过程的时延开销为目标,将边缘网关部署问题描述成一个二进制整数规划,同时设计了一种基于Q-learning强化学习的改进蜘蛛猴算法求解该二进制整数规划问题。仿真结果表明,相比于传统智能搜索算法,文章所设计的Q-learning改进蜘蛛猴算法能够以更高的求解和收敛效率求解边缘网关部署问题。The intelligent IoT platform for electrical equipment uses the edge gateway deployed on the power communication network’s router to capture and forward data from the supplier's data collection node for the achievement of efficient data interaction between platform and its suppliers.In this scenario,aiming at the huge production data and supplier’s data collection nodes,this paper describes the edge gateway deployment problem as a binary integer programming model to minimize the delay of the data capture and forwarding process.Meanwhile,this paper designs an improved spider monkey algorithm based on Q-learning reinforcement learning to solve this binary integer programming problem.The simulation results show that compared with three traditional intelligent searching algorithms,the spider monkey algorithm improved with Q-learning algorithm can obtain higher solving ability and convergence efficiency in solving the problem of edge gateway deployment of electrical equipment platform.

关 键 词：电工装备 边缘网关 蜘蛛猴算法 二进制规划 

分 类 号：TN91[电子电信—通信与信息系统]