摘要
针对可重入混合流水车间调度问题(RHFSP),提出一种协作蛙跳算法(CSFLA),以同时最小化最大完成时间和总延迟时间.给出了模因组的解质量和进化质量评价方法,根据进化质量确定最多两对模因组,在每对的两个模因组之间执行交换搜索次数和搜索能力的动态协作,并运用动态多邻域搜索(DMNS)和自学习过程改善算法性能.运用大量实例进行仿真实验,实验结果表明:CSFLA的新策略有效,且在多目标RHFSP优化方面具有较强的优势.
A cooperated shuffled frog-leaping algorithm(CSFLA)was proposed to minimize makespan and total tardiness for reentrant hybrid flow shop scheduling problem(RHFSP).A method for evaluating the solution quality and evolution quality of memeplexes was given.At most two pairs of memeplexes were determined according to the evolution quality,and a dynamic cooperation of exchanging search times and search ability were carried out between the two memeplexes of each pair.Dynamical multiple neighborhood search(DMNS)and self-learning process were added to improve the algorithm performance.Through extensive simulation experiments,experiment results show that the new strategies of the CSFLA are effective,which has promising advantages for the multi-objective RHFSP.
作者
雷德明
刘敬裕
LEI Deming;LIU Jingyu(School of Automation,Wuhan University of Technology,Wuhan 430070,China)
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2023年第5期125-130,共6页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(61573264)。
关键词
混合流水车间调度
蛙跳算法
多目标优化
可重入
进化质量
hybrid flow shop scheduling
shuffled frog-leaping algorithm
multi-objective optimization
reentrant
evolution quality
