主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2009, Vol. 17 ›› Issue (5): 68-74.

• 论文 • 上一篇    下一篇

炼钢连铸生产调度问题的两阶段遗传算法

李铁克, 苏志雄   

  1. 北京科技大学经济管理学院, 北京 100083
  • 收稿日期:2008-12-05 修回日期:2009-09-06 出版日期:2009-10-31 发布日期:2009-10-31
  • 作者简介:李铁克(1958- ),男(汉族),吉林长春人,北京科技大学经济管理学院,教授,博士生导师研究方向:先进制造管理与生产计划调度.
  • 基金资助:

    国家自然科学基金资助项目(70371057,70771008)

Two-Stage Genetic Algorithm for SM-CC Production Scheduling

LI Tie-ke, SU Zhi-xiong   

  1. School of Economics and Management, University of Science and Technology Beijing, Beijing 100083, China
  • Received:2008-12-05 Revised:2009-09-06 Online:2009-10-31 Published:2009-10-31

摘要: 将炼钢连铸生产过程抽象为混合流水车间,建立了0-1型混合整数线性规划调度模型。模型将严格连续浇注作为等式约束,并通过分段惩罚来平衡炉次的驻留时间。在对模型进行Benders分解的基础上,提出了将GA与LP结合的两阶段遗传算法。在算法设计中,提出了一种新的染色体编码来表示炉次设备指派与排序方案,给出了相应的遗传操作方法。算法的第一阶段通过最小化设备析取冲突来寻找高质量的种群,第二阶段通过求解线性规划模型来指导遗传算法的迭代过程。基于生产实际数据的仿真实验表明,该算法能够有效求解炼钢连铸生产调度问题。

关键词: 生产调度, 炼钢连铸, 遗传算法, 数学规划

Abstract: Steelmaking-continuous casting production process can be abstracted as a hybrid flow-shop.AO-1 mixed-integer linear programming model is established for this scheduling problem. In this model,no dead time inside the same cast at the last stage is treated as equality constraint,and graded penalty method is used to balance the sojourn times. Based on Benders'decomposition,a two-stage genetic algorithm combined GA and LP is proposed. In the algorithm design,a new chromosome encoding is used to represent the charge assignment and processing sequence solution,and genetic operations are given for this coding scheme. In the first stage,a high quality population by minimizing the weighted sum of overlapping time is found. And in the second stage the linear programming model to guide the iteration process is used. Finally,the result of simulation experiment with practical production data indicates that it is an efficient algorithm for this production scheduling problem.

Key words: production scheduling, steelmaking-continuous casting, genetic algorithm, mathematical programming

中图分类号: