改进多目标差分进化算法求解考虑供需平衡的热轧合同计划问题  

Improving multi-objective differential evolution algorithm for solving hot rolling order planning problem with supplydemand balance consideration

作　　者：赵伟[1] 谭清月 赵国栋[2] 张田 ZHAO Wei;TAN Qingyue;ZHAO Guodong;ZHANG Tian(Digital Intelligence Development Department,Angang Steel Co.,Ltd.,Anshan 114021,China;National Frontiers Science Center for Industrial Intelligence and Systems Optimization,Northeastern University,Shenyang 110819,China;Advanced Enterprise Performance Product Department,Beijing Yonyou Network Technology Company,Beijing 100120,China)

机构地区：[1]鞍钢股份有限公司数智发展部,辽宁鞍山114021 [2]东北大学工业智能与系统优化国家级前沿科学中心,辽宁沈阳110819 [3]北京市用友网络科技公司高级版企业绩效产品部,北京100120

出　　处：《冶金自动化》2023年第5期63-70,共8页Metallurgical Industry Automation

基　　金：国家自然科学基金重大项目(72192830,72192831)。

摘　　要：合理安排合同生产日期和满足后工序的合同需求可以保证合同生产的连续性,减少停机限产带来的资源浪费。针对一个计划周期内的合同排产问题,建立多目标优化模型,设计改进差分进化算法(three mutation strategies-multi-objective differential evolution,TMS-MODE),对不同个体采用自适应变异策略,从而兼顾算法的搜索深度和搜索广度。通过实际案例对模型和算法进行了验证,所提出的算法在中大规模问题上效果显著,其多样性、收敛性和均匀性都有较大优势,均优于目前常用的非支配排序遗传算法(non-dominated sorting genetic algorithms-Ⅱ,NSGA-Ⅱ)和基础的MODE。Reasonably arranging the order production dates and satisfying the order requirements of subsequent processes can ensure the continuity of order production and reduce the waste of resources caused by downtime and production restrictions.Aiming at the order scheduling problem within a planning cycle,a multi-objective optimization model was established,an improved differential evolution algorithm(three mutation strategies-multi-objective differential evolution,TMS-MODE)was designed,and an adaptive mutation strategy for different individuals was adopted to balance the search depth and breadth of the algorithm.The proposed algorithm was validated through practical cases,and it shows significant performance advantages over common NSGA-II and the basic MODE algorithm in terms of diversity,convergence,and uniformity in medium to large-scale scenarios.

关 键 词：热轧 合同计划 供需平衡 多目标差分进化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TG335.11[自动化与计算机技术—控制科学与工程]