引用本文:李冬,刘建昌,谭树彬,金阳,张彩金.改进蚁群算法在热精轧负荷分配优化中的应用[J].控制理论与应用,2014,31(8):1077~1086.[点击复制]
LI Dong,LIU Jian-chang,TAN Shu-bin,JIN Yang,ZHANG Cai-jin.Application of improved ant colony algorithm in load distribution optimization of hot finishing mills[J].Control Theory and Technology,2014,31(8):1077~1086.[点击复制]
改进蚁群算法在热精轧负荷分配优化中的应用
Application of improved ant colony algorithm in load distribution optimization of hot finishing mills
摘要点击 3006  全文点击 1598  投稿时间：2013-07-23  修订日期：2014-04-22
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2014.30771
  2014,31(8):1077-1086
中文关键词  热轧机  负荷分配  改进蚁群算法  优化
英文关键词  hot rolling mills  load distribution  improved ant colony algorithm  optimization
基金项目  国家自然科学基金资助项目(50974145).
作者单位E-mail
李冬* 东北大学 信息科学与工程学院 lidong_lu@163.com 
刘建昌 东北大学 信息科学与工程学院  
谭树彬 东北大学 信息科学与工程学院  
金阳 东北大学 信息科学与工程学院  
张彩金 中铝瑞闽铝板带有限公司装备能源部  
中文摘要
      本文尝试用改进的蚁群算法(IACA)求解热精轧机组负荷分配优化问题. 首先, 建立负荷分配优化的目标函数和约束条件. 为了避免蚁群算法(ACS)在加速收敛中出现停滞现象, IACA通过局部和全局信息素浓度更新、引入约束条件的蚂蚁视觉启发函数和基于轧制理论的智力启发函数对状态转移规则进行改进计算; 为了保证算法在迭代后期能够收敛, IACA动态更新状态转移规则中的阈值常数和局部信息素浓度挥发系数. 基于实际生产数据的仿真结果表明, IACA能够按照目标函数的要求进行合理的负荷分配, 且解的性能优于经验值; 与其他优化算法比较, IACA具有较快的求解速度和较高的求解精度.
英文摘要
      We employ the improved ant colony algorithm (IACA) to investigate the load distribution optimization problem of hot finishing mills. Firstly, objective functions and constraints of load distribution are established. In order to avoid the stagnation of ant colony system (ACS) in accelerating convergence process, IACA makes improvement on the calculation of state transition rule by updating the local and the global pheromone concentration and introducing constrained ant-visual heuristic function as well the intelligence heuristic function based on the rolling theory. In addition, to ensure the algorithm for converging in the last stage of iteration, the volatile coefficient of local pheromone concentration and the threshold constant in the state transition rule are all dynamically adjusted by IACA. Experimental results based on practical production data indicate that solutions given by IACA are able to make load distribution reasonable in accordance with requirements of objective functions, and to perform better than the empirical load distribution solution. In addition, IACA provides faster and more accurate solution than the other optimization methods.