分区编码HBGWO算法求解并行机分批调度问题  

Partition-encoding HBGWO algorithm for lot-sizing scheduling problem of unrelated parallel machines

作　　者：宋君乐 陶翼飞[1] 可晓东 丁小鹏 SONG Jun-le;TAO Yi-fei;KE Xiao-dong;DING Xiao-peng(Faculty of Mechanical and Electrical Engineering,Kunming University of Science and Technology,Kunming 650500)

机构地区：[1]昆明理工大学机电工程学院,云南昆明650500

出　　处：《机械设计》2022年第10期48-59,共12页Journal of Machine Design

基　　金：国家自然科学基金资助项目(51165014)。

摘　　要：考虑到不相关并行机调度问题中工件的批量与分批情况及机器调整时间,以最小化最大完工时间(Makespan)为优化目标,建立该NP-hard问题的数学规划模型,提出一种将工件批次、批量、机器分区的编码方式,设计了结合Logistic混沌映射策略、自适应变异算子、局部反向学习策略及基于狼群游走机制的混合生物地理学灰狼算法(HBGWO)对不同算例进行求解。对不同规模算例及实例进行仿真试验并与其他启发式算法求解结果进行了对比,对比结果证明了文中所提算法的有效性和优越性。In this article,with the lot-sizing and batching of jobs and the machine’s adjustment time in the scheduling of unrelated parallel machines,the mathematical programming model for NP-hard is set up with the focus on minimizing the maximum complete time(Makespan),and the partition-encoding method for batching,lot-size and assignment of machine is proposed.Efforts are made to combine the hybrid biogeography-based gray wolf(HBGWO)algorithm with the Logistic chaotic mapping strategy,the adaptive mutation operator,and the local reverse learning strategy;besides,based on the wolf-pack walking mechanism,different cases are solved.A series of scale cases and examples are subject to simulation experiments;the results are compared with those obtained from other heuristic algorithms.It’s shown that this algorithm is effective and superior.

关 键 词：不相关并行机 分批调度 灰狼算法 生物地理学算法 LOGISTIC映射 

分 类 号：TP278[自动化与计算机技术—检测技术与自动化装置]