摘要
分析了遗传系数学习时所遇到的问题 ,提出了鲁棒学习的概念和方法 ;以 2 0 5 0热轧精轧自然宽展为例 ,指出了现有精轧自然宽展模型的不足 ,提出了精轧宽度变化预测的统计模型 ,并基于遗传系数鲁棒学习方法进行了仿真研究。研究结果表明 ,文章提出的方法是行之有效的。
Primary problems in learning inherited parameters are analyzed and concept and methods of robust learning are presented.Disadvantages of the natural spread model for finish rolling are pointed out with natural spread of 2050mm hot finish rolling as an example.A statistical model for width-changing prediction for finish rolling is proposed,and a study on simulation is conducted based on the robust learning.The results show that the methods proposed in the paper prove to be effective.
出处
《宝钢技术》
CAS
2004年第1期55-58,61,共5页
Baosteel Technology
关键词
遗传系数
鲁棒学习
自然宽展
综合预报
Inherited parameter
Robust learning
Natural width spread
Integrated prediction
