求解混合流水车间调度的多目标优化算法  

Multi-objective optimization algorithm for hybrid flow shop scheduling problem

作　　者：谢光[1] 潘玉霞[1] 李俊青[2,3] 

机构地区：[1]三亚学院信息与智能工程学院,海南三亚572000 [2]聊城大学计算机学院,山东聊城252059 [3]东北大学流程工业自动化国家重点实验室,辽宁沈阳110819

出　　处：《计算机工程与设计》2018年第3期885-889,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61573178;61503170);国家重点实验室开放课题基金项目(PAL-N201602);海南省自然科学基金项目(617182)

摘　　要：为求解多目标混合流水车间调度问题,以最小化最大完工时间、提前惩罚量和滞后惩罚量为目标,提出一种改进的MOEA/D算法。采用基于排列的编码策略,设计两种局部搜索策略,提高算法求解性能。设计一种全局搜索交叉算子,给出一种种群更新策略,进一步提升解的分布均匀性。基于炼钢连铸实际生产数据,随机产生20个算例,与最新的两种多目标优化算法进行对比分析,验证了所提算法的优越性。To solve the multi-objective hybrid flow shop scheduling problems,an improved MOEA/D algorithm was presented.Three objectives were minimized simultaneously,i.e.,the minimization of the maximal completion time,earliness penalty,and tardiness penalty.A sequence-based coding mechanism was introduced.To increase the performance,two types of local search approaches were developed.A global search based crossover operator was presented.A population update mechanism was designed to further enhance the population diversity capabilities.Based on the steelmaking production reality,20 instances were randomly generated.Compared with two recently published efficient algorithms on the 20 instances,the superiority of the proposed algorithm is verified.

关 键 词：混合流水车间调度 多目标优化算法 关键路径 全局搜索 局部搜索 

分 类 号：TP278[自动化与计算机技术—检测技术与自动化装置]