摘要
针对基于传统模型的方法难以在线优化磨矿过程回路设定值的问题,提出了基于案例推理与强化学习的运行指标优化方法,建立基于自回归神经网络的Q函数模型,并应用案例推理更新模型连接权值,实现了磨矿过程关键参数的实时优化。
It is difficult to optimize the setpoint of ore grinding process loop online base on the method of conventional model.Aiming at this problem,an optimizing method of operation index based on case-based reasoning and reinforcement learning is proposed.The real-time optimization for key pa-rameters of grinding process is realized through establishing Q function model based on autoregression neural network and updating model connection weights by apply ing case-based reasoning.
作者
徐凯
罗赛
陈洪彬
XU Kai;LUO Sai;CHEN Hong-bin(Angang Group Guanbaoshan Mining Industry Co.,Ltd;Shenyang Automation Institute of Chinese Academy ofSciences;Robotics&Intelligent Manufacturing and Innovation Institute of Chinese Academy of Sciences;Angang Group Mining Industry Design&Research Institute,Anshan 114044)
出处
《冶金设备管理与维修》
2022年第1期5-6,7,8,共4页
Metallurgical Equipment Management and Maintenance
关键词
案例推理
强化学习
Q函数
设定值优化
磨矿过程
Case-based reasoning
reinforcement learning
Q function
setpoint optimization
grinding process
