Abstract: The development of mass customization has led to an increasing variety of products. In order to meet diversified customer needs, a mixed model assembly line that allows different models to be assembled on a same production line has emerged. Tens of thousands of parts need to be delivered to the assembly line in time, which brings enormous challenge on part supply faced by manufacturing enterprises. In order to meet part requirements of mixed model assembly line, frequent deliveries with small batch are becoming a trend recently. A new material supermarket is introduced to promote just-in-time part supply for mixed model assembly line.
The part feeding problem refers to the logistics process of delivering the matched parts to the corresponding stations on time according to the consumption of parts. Combined with the new part-feeding mode with material supermarket, this paper proposes a just-in-time part feeding method that is different from traditional periodic part feeding. According to the problem description and related assumptions, a mixed integer programming model with the goal of minimizing work-in-process inventory is constructed to determine the optimal loading parts and schedule of each tour under the constraints of tow train capacity and preventing part shortage. A heuristic algorithm is designed to solve the large-scale problems after considering the continuity and complexity of modern production. The heuristic algorithm initially arranges the part type and the start time of each tour according to the calculated shortage point. Once the interval time of two consecutive tours is overlapped, the start time needs to be coordinated by backward pushing.
To evaluate the computational performance of the proposed heuristic algorithm, problems of different scales are employed in the numerical analyses. The results obtained from the proposed heuristic algorithm are compared with results solved by Cplex and the genetic algorithm. The comparison demonstrates that the proposed heuristic can get all the optimal solutions for the small-sized problems, and it outperforms genetic algorithm in terms of both solution quality and solution time. Furthermore,the just-in-time part feeding method is analyzed and compared with the widely used periodic part feeding under different tow train capacities as well as different material requirements. The results show that the just-in-time part feeding method proposed in this paper makes the delivery time as close to the shortage time of each part as possible. Compared with the widely used periodic part feeding method, it can more effectively reduce the inventory of the workstations, the occupation of workstation space and production costs. At the same time, the capacity of the tow train also has an impact on the average inventory, and it needs to be selected reasonably according to the shortage time to reduce the average inventory.
Although there have been many researches on the part feeding problem for mixed model assembly lines, most of the existing researches focus on periodic part-feeding. The research in this paper provides the just-in-time scheduling for the tow train which travels between material supermarket and assemble line. It is of great significance to enrich and expand the research field of part supply process based on material supermarket. In the future, we will take more tow trains into consideration while this paper only allows one tow train which simplifies the problem.

Key words: mixed model assembly line, material supermarket, just-in-time part supply, vehicle scheduling

摘要： 针对混流装配线物料配送优化问题,结合新型的物料超市配送模式,提出了区别于传统周期性配送的准时化配送方法。首先,根据问题描述和相关假设构建了以线边库存最小化为目标的混合整数规划模型。然后,考虑到实际生产的连续性和复杂性,为求解大规模问题,设计出适用于问题的启发式算法,并与遗传算法进行比较,启发式算法能在更短时间内获得满意解。最后,在不同牵引拖车容量下,通过实例分析,与广泛应用的周期性配送进行比较,算例求解结果表明,准时化配送模型能够获得更低的库存水平,减少工作站的空间占用。

关键词: 混流装配线, 物料超市, 准时化配送, 车辆调度