摘要
通过在离散生产线上应用RFID(radio frequency identification, RFID)技术用以实现产线动态感知功能,研究基于深度强化学习(deep reinforcement learning, DRL)算法的生产线动态感知决策方法。对于具有离散生产作业的多机生产线,提出利用数据采集量(DCV)和数据采集水平(QCL)两类性能指标对产线智能化程度进行精准量化,建立基于马尔可夫链的生产线动作性能评估模型和基于马尔可夫决策过程的生产线动态感知决策模型。利用DRL算法对问题进行近似求解,获得了有效的生产线动态感知策略。实验结果表明,所提出的生产线动态感知决策模型可以有效地帮助企业选择与生产内容高度匹配的RFID硬件和合理有效的硬件安装位置,提高生产线的数据采集能力,打造智能产线样板。
By selecting the key enabling technology RFID for smart manufacturing perception layer and introducing it into the discrete production line of SMEs for work in process(WIP)quality control,we study the dynamic perception decision method of production line based on deep reinforcement learning(DRL)algorithm.For a multi-machine production line with discrete production operations,two types of performance indicators,data collection volume(DCV)and data collection level(QCL),are proposed to accurately quantify the intelligence of the production line,and a Markov chain(MC)based production line action performance evaluation model and a Markov decision process(MDP)based production line dynamic perception decision model are established.The DRL algorithm is used to approximate the problem and obtain an effective dynamic perception strategy for the production line.The experimental results show that the proposed production line dynamic sensing decision model can effectively help enterprises to select RFID hardware that highly matches the production content and reasonable and effective hardware installation locations to improve the data collection capability of the production line.
作者
黄松勇
王贤琳
鄢威
张豪
HUANG Songyong;WANG Xianlin;YAN Wei;ZHANG Hao(Green Manufacturing Engineering Research Institute,Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,China;Hubei Provincial Key Laboratory of Mechanical Transmission and Manufacturing Engineering,Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,China;Key Laboratory of Metallurgical Equipment and Its Control,Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,China)
出处
《组合机床与自动化加工技术》
北大核心
2023年第8期176-182,共7页
Modular Machine Tool & Automatic Manufacturing Technique
基金
国家自然科学基金资助项目(51975432)。
关键词
智能制造
离散生产
动态感知决策
RFID系统
深度强化学习
smart manufacturing
discrete production line
dynamic sensing decision
RFID system
deep reinforcement learning
