基于改进型NSGAⅡ的织造车间多目标大规模动态调度  

Multi-objective large-scale dynamic scheduling for weaving workshops based on improved NSGAⅡ

作　　者：沈春娅 雷钧杰 汝欣 彭来湖[1,2] 胡旭东[1,2] SHEN Chunya;LEI Junjie;RU Xin;PENG Laihu;HU Xudong(School of Mechanical Engineering and Automation, Zhejiang Sci-Tech University, Hangzhou, Zhejiang 310018, China;Key Laboratory of Modern Textile Machinery & Technology of Zhejiang Province,Zhejiang Sci-Tech University, Hangzhou, Zhejiang 310018, China)

机构地区：[1]浙江理工大学机械与自动控制学院,浙江杭州310018 [2]浙江理工大学浙江省现代纺织装备技术重点实验室,浙江杭州310018

出　　处：《纺织学报》2022年第4期74-83,共10页Journal of Textile Research

基　　金：浙江省公益技术研究计划项目(LGG21E050024);浙江省重点研发计划项目(2019C01038);浙江省博士后科研项目特别资助项目(ZJ2020004);浙江理工大学科研启动基金项目(18022224-Y)。

摘　　要：织造车间调度规模普遍在300台织机、1000个织轴以上,遗传算法搜索极易陷入局部最优,针对传统动态调度机制在织造插单、打样等复杂生产场景中适应性不强的问题,提出一种改进NSGAⅡ算法。从织造多织机、多织轴、多产品的大规模调度出发,基于织造和穿经之间独特的逆工序调度关系,构建以逾期损失、最大完工时间和织机空闲时间均最小为目标的织造多目标大规模调度模型。通过改进启发规则的编码方式缩小解空间,设计了一种局部和全局关联优化的贪婪进化算子,避免算法寻优陷入局部最优;并提出基于支配关系评价的动态调度机制,优化算法在生产中动态响应机制差,抗扰动性不高的不足。验证实验证明,改进NSGAII算法在织机调度规模为500台、4000个织轴时,调度能力仍优于其他算法。As the number of looms exceeds 300 with more than 1 000 weaver′s beams in the weaving workshop, the genetic algorithm is easy to fall into local optimal solution when solving such large-scale scheduling problems, and the traditional dynamic scheduling mechanism is not adaptable enough for complex production scenarios such as order insertion and proofing. An improved NSGAⅡ algorithm was proposed in the paper. Considering the facts that the scheduling of a large-scale weaving workshop involves large numbers of looms, weaver′s beams and products, and the unique inverse process scheduling relationship between weaving and drawing-in, a multi-objective large-scale scheduling model for weaving was constructed, aiming at the minimization of overdue loss, makespan, and idle time of loom. The encoding of heuristic rules was improved to reduce the solution space, and a greedy evolution operator was used in local and global correlation optimization to avoid falling into local optimization. A dynamic scheduling mechanism based on dominance relationship evaluation was adopted to improve the poor dynamic response mechanism and low ability against disturbance during production. Experiments show that the scheduling ability of the algorithm remains superior over other algorithms in a situation where there are 500 looms with 4 000 weaver′s beams in a weaving workshop.

关 键 词：织造车间智能调度 NSGAⅡ 多目标优化 大规模调度 动态调度 启发规则 

分 类 号：TS111.8[轻工技术与工程—纺织材料与纺织品设计]