基于空间蚁群算法的混流装配线组批排序方法  

Mixed-Model Assembly Line Batch Sequencing Method Based on Space Ant Colony Algorithm

作　　者：蒙秋男[1] 王丹迎[1] 

机构地区：[1]大连理工大学管理与经济学部,辽宁大连116024

出　　处：《工业工程与管理》2016年第3期94-103,共10页Industrial Engineering and Management

基　　金：国家自然科学基金资助项目(71172137;61034003);国家科技支撑计划项目(2015BAF08B02)

摘　　要：针对混流装配线按工作日历调度过程中,组批生产导致订单准时交付率差、加班时间长,以及缺料等干扰因素导致完工周期延长等问题,以零部件配套、按工作日历组批为约束,最小化完工周期、提前/拖期时间以及加班时间为多目标,建立工件排序数学优化模型。验证以完工周期最短为目标的流水线工件最优排序亦具有V型特征。提出空间蚁群权重设计方法,将蚂蚁沿不同的权重向量寻优,提高算法的全局搜索能力;并根据当前Pareto解在各权重子空间的分布情况动态调整各子空间的蚂蚁数量,避免陷入局部最优。通过与文献算法对比,验证空间蚁群算法具有良好的优化性能,并通过实例验证了排序方法的有效性。A mathematical sequence optimization model with constraints of components kitting and batch production based on shop calendar,aiming at minimize total completion time,total earliness/tardiness time and overtime is established.The proposition that the optimal job sequence considering shortest total completion time also has a characteristic of ' V' is proved.A space ant colony algorithm with a new weight determination mechanism is designed to solve this problem,in which ant searched solutions along a randomly generated weight direction in its subspace,thus improving the global searching capability.Besides,ants are assigned to the corresponding sub-space according to the distribution of current Pareto solutions,this efficiently avoids algorithm falling into local optimum.Compared to the results of the algorithm in literature,the space ant colony algorithm has better optimizing performance.A case study is also given to validate the proposed sequencing method.

关 键 词：混流装配线组批排序 空间蚁群算法 零部件配套 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]