引用本文:李 勇,刘建昌,王 昱.改进权重自适应GA及冷连轧轧制规程多目标优化[J].控制理论与应用,2009,26(6):687~693.[点击复制]
LI Yong,LIU Jian-chang,WANG Yu.An improved adaptive weight approach GA for optimizing multi-objective rolling schedules in a tandem cold rolling[J].Control Theory and Technology,2009,26(6):687~693.[点击复制]
改进权重自适应GA及冷连轧轧制规程多目标优化
An improved adaptive weight approach GA for optimizing multi-objective rolling schedules in a tandem cold rolling
摘要点击 1524  全文点击 909  投稿时间:2008-04-24  修订日期:2008-11-18
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/j.issn.1000-8152.2009.6.021
  2009,26(6):687-693
中文关键词  改进权重自适应方法  遗传算法  冷连轧  轧制规程  多目标优化
英文关键词  improved adaptive weight approach  genetic algorithm  tandem cold rolling  rolling schedules  multiobjective optimization
基金项目  高等学校博士学科点专项科研基金资助项目(20060145025); 辽宁省自然科学基金资助项目(20052033).
作者单位E-mail
李 勇 东北大学 流程工业综合自动化教育部重点实验室, 辽宁 沈阳 110004 liyong.neu@gmail.com 
刘建昌 东北大学 流程工业综合自动化教育部重点实验室, 辽宁 沈阳 110004 liujianchang@ise.neu.edu.cn 
王 昱 沈阳航空工业学院 自动化学院, 辽宁 沈阳 110136 wangyu.siae@gmail.com 
中文摘要
      针对聚合多目标优化方法的权重难以确定的问题, 提出了一种改进的权重自适应方法, 并以遗传算法为基础对冷连轧轧制规程进行多目标优化. 首先, 结合某冷轧厂实际的轧制规程优化过程, 选取等功率裕量、轧制能耗及带钢打滑概率作为优化目标, 建立了冷连轧轧制规程的多目标优化模型. 然后将改进的权重自适应遗传算法(GA)应用于不同规格的带钢轧制规程多目标优化中, 结果表明, 与实际应用的轧制规程相比, 该方法有效的降低了3个目标函数的值; 与权重自适应GA相比, 改进的权重自适应GA的针对性更强, 同时重要性高的目标收敛速度更快.
英文摘要
      To select correct weighting factors for the weight-sum multi-objective method, we propose an improved adaptive weight-selection approach. On the basis of the genetic algorithm(GA), this approach is applied to optimize the multi-objective rolling schedules in a tandem cold rolling. In the optimization process of rolling schedules, the power distribution, the rolling energy consumption and the slip rate are selected as objective functions from them the multiobjective model of rolling schedules is established. Applying the improved adaptive weight approach GA to optimize rolling schedules for strips with different specifications, we reduce the values of the above three objective functions simultaneously, in comparison with the conventional rolling schedules. It also provides better pertinence and faster convergence for objects of higher priority than those of the adaptive weight approach GA.