钢铁冷轧多机组排产算法设计  

Design of multi-process scheduling algorithm for steel cold rolling

作　　者：沈春锋[1] 杨海飞 SHEN Chunfeng;YANG Haifei(Academy of Industrial Internet/Big Data Center,Shanghai Baosight Software Co.,Ltd.,Shanghai 201203,China)

机构地区：[1]上海宝信软件股份有限公司工业互联网研究院/大数据中心,上海201203

出　　处：《冶金自动化》2021年第2期16-22,共7页Metallurgical Industry Automation

基　　金：上海市优秀学术/技术带头人计划资助项目(18XD1420600);上海市科委社会发展领域科研项目(19DZ1209005)。

摘　　要：针对钢铁冷轧生产,以柔性车间调度问题(flexible job-shop scheduling problem,简称FJSP)为基础,总结出一个冷轧多机组排产问题,并提出一种通用高效的多机组排产算法。该算法主要由启发式的初始化、邻域搜索与跳跃搜索模块组成,各模块可根据具体约束条件,灵活配置相应的启发式算子。实验结果表明,该算法相对于传统的变邻域搜索能找到更优的解,且在数据规模增大时保持很快的求解速度,具有高灵活度、高效率的优点。该算法已在某钢铁企业冷轧产线成功上线,取得了良好的效果。Aiming at the cold rolling production of iron and steel,a multi-process scheduling problem in steel cold rolling production was summarized based on FJSP,and a general and efficient multiprocess scheduling algorithm was proposed. The algorithm comprises initialization,neighborhood search and jumping search module in a heuristic way. Flexible heuristic operators can be configured for each module to adapt different constraints. Experiment results show that the proposed method can find a superior solution compared to traditional variable neighborhood search,and remain fairly fast solution speed while increasing data size. The proposed algorithm has the advantages of high flexible and computational efficient,and has been successfully put into use in a cold rolling line of an iron and steel enterprise and achieved good effect.

关 键 词：冷轧 排产算法 遗传算法 变邻域搜索 启发式算法 

分 类 号：TG335.12[金属学及工艺—金属压力加工] TP18[自动化与计算机技术—控制理论与控制工程]