基于混合蚁群分布估计算法的可重入流水车间调度  

Reentrant Flow Shop Scheduling Based on Hybrid Ant Colony Distribution Estimation Algorithm

作　　者：钟臻怡 杨家荣[1] 吕伟[1] ZHONG Zhenyi;YANG Jiarong;LV Wei(Central Academe,Shanghai Electric Group Co.,Ltd.,Shanghai 200070)

机构地区：[1]上海电气集团股份有限公司中央研究院,上海200070

出　　处：《计算机与数字工程》2021年第5期863-868,895,共7页Computer & Digital Engineering

基　　金：上海市科研计划项目“基于多维度矩阵的智能排程技术”(编号:18DZ1100702)资助。

摘　　要：在半导体晶圆制造过程中,调整设备相关参数来提高晶圆质量的同时会造成换模时间的增加,从而影响生产效率。并且考虑到半导体晶圆制造可重入的特性,论文对带换模约束的可重入流水车间进行了研究,提出了混合蚁群分布估计算法。首先对该问题进行了问题描述,并以系统makespan最小化为目标建立数学规划模型,随后将分布估计算法融入蚁群算法,以提高蚁群算法的全局搜索能力。最后实验结果验证了所提出算法的有效性。During the process of semiconductor manufacturing,the relevant parameters of machines are adjusted to improve wafer quality,but this increases the setup time,which reduces production efficiency.In this thesis,the characteristic of reentrance is considered at the same time.The paper studies the reentrant hybrid flow shop problem with setup constraints and puts forward a hybrid ant colony algorithm with estimation of distribution algorithm.Firsly the scheduling problem domain is described and mathe⁃matical programming formulations are put forward with the objective of minimizing makespan of the system.Then the estimation of distribution algorithm is used to improve the global search ability of the ant colony algorithm.Finally,simulation experiments dem⁃onstrate the effectiveness of the proposed algorithm.

关 键 词：换模 调度 蚁群算法 分布估计算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]