The problem of sequencing assembly lines consists of determining the order in which a given set of products is launched down the line. Since individual products may require different parts in different quantities, the production sequence has a big influence on line-side inventory. Classically, sequences are often optimized with the goal of attaining level schedules, i.e., the part demand should be smooth during the planning horizon. However, this approach does not necessarily work well if parts are delivered at discrete points in time in bulk quantities. In this paper, we consider a production system where bins of parts are delivered periodically by a tow train from a central depot at fixed times. Due to the limited space at the assembly line, the maximum number of bins in stock at any time at any station should be minimal. We propose an exact solution method based on combinatorial Benders decomposition as well as bounding procedures and heuristics for this problem. The algorithms are shown to perform well both on instances from the literature and on new data sets. We also investigate whether classic level scheduling methods are effective at reducing line-side stock in an assembly system supplied by tow train, and to what degree line-side stock can be traded off for more frequent deliveries.

中文翻译：

---

对装配线进行排序以促进同步的及时零件供应

对装配线进行排序的问题包括确定一组给定产品在生产线上下线的顺序。由于单个产品可能需要不同数量的不同零件，因此生产顺序对生产线侧库存有很大影响。传统上，序列通常以达到水平计划为目标进行优化，即零件需求在计划范围内应该是平滑的。但是，如果零件在离散时间点以大批量交付，则此方法不一定有效。在本文中，我们考虑了一个生产系统，其中零件箱在固定时间由中央仓库的拖车定期交付。由于装配线上的空间有限，任何时间任何站点的最大库存箱数都应该是最小的。我们针对这个问题提出了一种基于组合 Benders 分解以及边界程序和启发式的精确求解方法。这些算法在文献实例和新数据集上都表现良好。我们还研究了经典的水平调度方法在减少由拖车提供的装配系统中的线侧库存方面是否有效，以及线侧库存在多大程度上可以权衡更频繁的交付。