引用本文:汪洋,吴子轩,李铁克,王柏琳.考虑柔性机器检修的钢管热轧批量调度方法[J].控制理论与应用,2019,36(9):1536~1544.[点击复制]
WANG Yang,Wu Zi-xuan,Li Tie-ke,Wang Bai-lin.Methods of hot-rolled batch scheduling for seamless steel tube with flexible machine maintenance[J].Control Theory and Technology,2019,36(9):1536~1544.[点击复制]
考虑柔性机器检修的钢管热轧批量调度方法
Methods of hot-rolled batch scheduling for seamless steel tube with flexible machine maintenance
摘要点击 1422  全文点击 672  投稿时间：2018-01-02  修订日期：2019-05-22
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2019.80006
  2019,36(9):1536-1544
中文关键词  机器柔性检修  热轧批量调度  启发式算法  无缝钢管  序列相关
英文关键词  flexible machine maintenance  hot-rolled batch scheduling  heuristic algorithm  seamless steel tubes  sequence-dependent
基金项目  国家自然科学基金资助项目(71231001, 71701016)；北京市自然科学基金项目(9174038)；教育部人文社会科学研究青年基金项目(17YJC630143)；中央高校基本科研业务费资助项目(FRF-BD-16-006A）
作者单位E-mail
汪洋 北京科技大学东凌经济管理学院 fran_wangyang@qq.com 
吴子轩 北京科技大学东凌经济管理学院  
李铁克 北京科技大学东凌经济管理学院  
王柏琳* 北京科技大学东凌经济管理学院  
中文摘要
      本文从无缝钢管生产管理中提取并定义了周期性机器柔性检修环境下的钢管热轧批量调度问题，针对无缝钢管热轧阶段的生产特点，将其抽象为一类考虑序列相关设置成本和机器柔性检修的单机调度问题，建立了以最小化机器闲置时间和机器调整时间为优化目标的数学模型。分析闲置时间和检修时点的关系，证明了闲置时间最小化性质，结合问题特征设计了两阶段启发式算法。算法第一阶段采用最小轧机调整时间规则获取具有最小机器调整时间的初始批量轧制序列，第二阶段对初始轧制序列进行全局寻优搜索。基于实际生产数据设计了多种问题规模的对比实验，实验结果表明模型和算法对求解该类问题具有较好效果。
英文摘要
      A hot-rolled batch scheduling problem with periodic machine maintenance for steel tube was extracted and defined from the production management of seamless steel tube. According to the characteristic of the hot-rolled process, this problem was abstracted into a single machine scheduling problem with sequence-dependent setup cost and flexible machine maintenance, and a mathematical model was built with the objectives to minimize idle time and adjustment time of the machine. By analyzing the relationship between the idle time and the location of the maintenance, a property with respect to the idle time minimization was given. A two-stage heuristic algorithm was presented based on this property and other characteristics of the problem. At the first stage of the algorithm, a rule to minimize the rolling mill setup times was used, and the initial sequence with the smallest machine adjustment time was produced. At the second stage, a global optimization search was performed based on this initial rolling sequence to optimize the overall goal value. With the actual production data, computational experiment were carried out on the problems with various problem size, and experimental results showed that the model and algorithm had obvious effect on solving this problem.