Performance Evaluation and Optimization of Block Production Systems

为了提高生产的灵活性，最近引入了一种装配线的变体，其中工人比工位更少，工人步行到相邻的工位继续生产一件物品。从它的形式来看，这个系统被称为U型生产系统。U型生产系统实际上是通过改变工作分配来缓冲自己免受变化的影响，这允许系统以非常低的在制品库存(WIP)水平运行。对工人的控制有几个版本。一种是固定分配，每台机器只属于一名工人。另一种是自由浮动系统，在这种系统中，任何工人都可以在任何机器上工作。在这项研究中，我们提出了U型生产系统的排队模型，该模型包括固定分配系统和自由浮动系统作为特例。通过现场工作，我们观察到产量、在制品和工人数量是U型生产系统中的重要因素。然后我们考虑了以下优化问题：(1)在期望吞吐量约束下，分配固定数量的工人以最小化在制品(WIP)。(2)在期望吞吐量约束下，求出工人总数并分配工人，以最小化在制品和工人的总成本。由于传统排队理论很难解决实际规模的问题，我们将样本路径优化的思想应用到上述优化问题中。分析结果表明，变异系数对U型生产系统的最优结构有很大影响。

To increase flexibility of production, a variation of the assembly line has recently been introduced in which there are fewer workers than stations and workers walk to adjacent stations to continue work on an item. The system is called a U-shaped production system from its form. The U-shaped production system is, in effect, buffering itself against variation by altering the work assignments, which allows the system to operate with very low levels of work-in-process inventory (WIP). The control of the workers has some versions. One is Fixed Assignments in which each machine belongs to exactly one worker. Another is Free-floating System in which any worker can work on any machine. In this research, we have proposed the queueing model for U-shaped production systems, which includes both the fixed assignments and the free-floating systems as a special case. By the field work, we have observed the throughput, WIP, and the numbers of workers are important factor in U-shaped production systems. Then we have considered following optimization problems:(1) Under the restriction of the desired throughput, assign the fixed number of workers in order to minimize WIP(2) Under the restriction of the desired throughput, obtain the total number of workers and assign the workers in order to minimize the total cost for WIP and workersSince it is very difficult to solve the practical size of the problems using the traditional queueing theory, we apply the idea of sample-path optimization to the above optimization problems. The analytical results show that the coefficient of variation has a great influence on the optimal structure in the U-shaped production systems.