IMOPSO-Ⅱ求解多目标动态多机协作作业车间调度研究OA
IMOPSO-Ⅱ for solving multi-objective dynamic hybrid job-shop scheduling with multiprocessor tasks
为满足离散制造业个性化生产需求并应对突发事件,设计了一种紧急订单插入的多机协作作业车间调度模型.模型将调度过程分为原始调度和重调度两个阶段,以最小化最大完工时间和总拖延时间为目标进行求解.提出改进的多目标粒子群(Improved Multi-Objective Particle Swarm Optimization,IMOPSO-Ⅱ)算法,结合滚动窗口技术将动态区间转化为多个静态区间,选用改进优先操作交叉策略与多轮变异丰富种群,通过快速非支配排序和拥挤度计算选取优秀粒子,采用外部档案策略进一步保留优秀基因.生成5-Job和10-Job算例进行多轮实验,与结合遗传算法(Genetic Algorithm,GA)和非支配排序遗传算法(Non-dominated Sorting Genetic Algorithm,NSGA-Ⅱ)的 GA-NSGA-Ⅱ 算法相比,IMOPSO-Ⅱ算法在解的适应度值和C值上表现更优,验证了其在动态多机协作作业车间调度中的有效性.
A hybrid job-shop scheduling model with multiprocessor tasks was designed to meet personalized production needs in discrete manufacturing and handle emergencies.The model divided the scheduling process into initial scheduling and rescheduling stages,aiming to minimize the makespan and total tardiness.Improved Multi-Objective Particle Swarm Optimization(IMOPSO-Ⅱ)algorithm combined with a rolling window technique was proposed to convert dynamic intervals into static ones.The algorithm used an improved crossover strategy and multi-round mutation to enhance population diversity.Fast non-dom-inated sorting and crowding distance were applied to select particles,while an external archive preserved superior genes.Multiple experiments using 5-Job and 10-Job instances were conducted.Compared with the GA-NSGA-Ⅱ algorithm,IMOPSO-Ⅱ algorithm showed better performance in fitness values and C-metric,confirming its effectiveness in dynamic hybrid job-shop scheduling model with multiprocessor tasks.
樊坤;莫雅婧;瞿华;王君岩
北京林业大学经济管理学院,北京 100083北京林业大学经济管理学院,北京 100083北京林业大学经济管理学院,北京 100083北京林业大学经济管理学院,北京 100083
信息技术与安全科学
作业车间调度紧急订单多目标优化粒子群算法
job-shop schedulingurgent ordersmulti-objective optimizationparticle swarm optimization algorithm
《现代制造工程》 2026 (4)
7-15,77,10
教育部人文社科基金项目(21YJA630012)北京林业大学中央高校基本科研业务费专项资金项目(2023SKY06)
评论