一种考虑多种可选坯型的炉次批量计划优化方法研究与应用  

Research and application of an optimization method for charge batch planning considering multiple available size of steel billet

作　　者：王悦晓 施灿涛[1] 史忠轩 杨星 李志伟[2] 陈红雨[2] WANG Yuexiao;SHI Cantao;SHI Zhongxuan;YANG Xing;LI Zhiwei;CHEN Hongyu(Center of Industry Intelligence,China Metallurgical Industry Planning and Research Institute,Beijing 100711,China;Manufacturing Management Department,Ningbo Iron and Steel Co.,Ltd.,Ningbo 315807,China)

机构地区：[1]冶金工业规划研究院工业智能研究中心,北京100711 [2]宁波钢铁有限公司制造管理部,浙江宁波315807

出　　处：《冶金自动化》2023年第2期48-56,共9页Metallurgical Industry Automation

摘　　要：通过与钢铁企业实际生产相结合,面向订单坯料设计时多种可选坯型的特殊场景,充分考虑了炉次计划编制过程中所涉及的客户特殊要求、生产工艺约束、设备约束等条件,以降低钢水冶炼成本、减少无委托钢坯库存、提升客户服务水平为目标,基于约束理论建立了数据模型。通过挖掘模型特点,提出了一种基于贪婪算法和回溯法、约束传播方法混合的约束满足问题的求解方法。该算法与传统人工编制计划相比,计划编制效率平均提升2897倍,钢包利用率平均提高0.56%,人工调整率平均为2.38%。试验结果表明,本文所提供的炉次批量计划优化方法有效解决了人工编制计划模式下的效率低下、人工经验依赖程度高、重复性工作量大等问题。To reach the goal of cost reduction in steel making and stock and enhancing service level,a data model was established based on binding theory,combining with actual production of iron and steel enterprises,fully considering the multi selection of size of steel billet during designing suitable steel billet for order and special demand from customer,constraint of production process as well as constraint of devices during charge planning formulation.Comparing to manual batch planning,this method proposed by mining the characteristics of the model boosts the planning efficiency by 2897 times and raises availability of steel ladle by 0.56%and average manual adjusting rate is 2.38%,which proves that the proposed mixed solution involving greedy algorithm,backtracking algorithm as well as constraint propagation algorithm under constraints that improves the efficiency of manually planning and minimize the influence of dependency of manual experience as well as the workload.

关 键 词：智能决策 炉次批量计划 多种可选坯型 约束满足 热轧板带 

分 类 号：TF341[冶金工程—冶金机械及自动化]