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 ›› 2024, Vol. 32 ›› Issue (8): 95-106.doi: 10.16381/j.cnki.issn1003-207x.2021.2590cstr: 32146.14.j.cnki.issn1003-207x.2021.2590

• • 上一篇    下一篇

加工时间不确定的炼钢-连铸区间多目标优化调度

李铁克1,2,苏艺璇1,2,张文新1,2,王柏琳1,2()   

  1. 1.北京科技大学经济管理学院, 北京 100083
    2.钢铁生产制造执行系统技术教育部工程研究中心, 北京 100083
  • 收稿日期:2021-12-13 修回日期:2022-04-28 出版日期:2024-08-25 发布日期:2024-08-29
  • 通讯作者: 王柏琳 E-mail:wangbl@ustb.edu.cn
  • 基金资助:
    国家自然科学基金项目(72301026);教育部社科研究基金规划项目(23YJA630090)

Interval Multi-objective Optimal Scheduling for Steelmaking-continuous Casting with Processing Time Uncertainty

Tieke Li1,2,Yixuan Su1,2,Wenxin Zhang1,2,Bailin Wang1,2()   

  1. 1.School of Economics and Management, University of Science and Technology Beijing, Beijing 100083, China
    2.Ministry of Education Engineering Research Center of MES Technology for Iron & Steel Production, Beijing 100083, China
  • Received:2021-12-13 Revised:2022-04-28 Online:2024-08-25 Published:2024-08-29
  • Contact: Bailin Wang E-mail:wangbl@ustb.edu.cn

摘要:

加工时间不确定是炼钢-连铸生产中普遍存在且具有代表性的一类不确定因素。针对其调度问题,采用三参数区间数描述加工时间不确定性信息,建立了以炉次总等待时间和浇次开浇时间提前/拖期总惩罚最小化为优化目标的区间多目标混合整数规划模型,并结合问题特征,设计了一种基于分类进化策略的改进快速非支配排序区间多目标遗传算法。在算法中,融合区间数相关操作提出了区间意义下的逆序并行倒推解码法和基于机器选择规则的种群初始化混合策略,设计了基于个体拥挤距离的交叉变异算子分类进化策略,并提出对种群中重复个体执行再变异操作,以维持种群多样性。基于实际生产数据的仿真实验验证了区间多目标优化算法在求解质量和求解效率方面的有效性。

关键词: 炼钢-连铸, 生产调度, 加工时间不确定, 区间多目标优化, 遗传算法

Abstract:

Due to the complex physical changes and chemical reactions in the process of the steelmaking-continuous casting (SCC), the uncertainty of processing time is a common and representative uncertainty factor. Therefore, it is necessary to consider the uncertain processing time before scheduling, to enhance the robustness of schedules and reduce the repair frequency of dynamic scheduling.For SCC scheduling problem with processing time uncertainty, the processing time is described by a three-parameter interval. A multi-objective optimization model with interval-valued is established to minimize the total waiting time and the total earliness/tardiness of casting time. To solve this problem, an improved fast elitist non-dominated sorting genetic algorithm (NSGAII+) based on a classification evolution strategy is presented.Firstly, a decoding scheme considering the reverse order and a hybrid population initialization based on machine rules are proposed combining interval number operation. Then a classification evolution strategy is adopted to determine the crossover and mutation operators according to the crowding distance. The re-mutation of repeated individuals is proposed to maintain the diversity of the population. Finally, the results of the experiments based on actual SCC production data shows the effectiveness of the proposed NSGAII+ in solving quality and efficiency.Note that if the upper and lower limits and intermediate parameter of all three parameter intervals are the same, the problem is transferred into a static scheduling based on standard processing time. If two of the upper and lower limits and intermediate parameters of all three parameter intervals have the same value, it will degenerate into a two-parameter interval number problem. Thus, the model and algorithm proposed in this paper are also applicable to the above two problems.

Key words: steelmaking-continuous casting, production scheduling, processing time uncertainty, interval multi-objective optimization, genetic algorithm

中图分类号: