Comparison between two types of inventory targets under variability of a semiconductor supply chain
Comparison between two types of inventory targets under variability of a semiconductor supply chain
复制标题
半导体供应链变化下两种库存目标的比较
DOI：
10.11648/j.ijber.s.2014030601.21
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
Geraldine Yachi
中科院分区：
文献类型：
--
作者：
Kenichi Nakashima;Thitima Sornmanapong;Hans Ehm;Geraldine Yachi
随着半导体行业创新的迅速发展，供应链流程并没有跟上。半导体供应链的可变性增加了，也变得更加复杂。这有助于准确预测需求并设定库存目标。需求和供给越来越具有随机性和非平稳性。库存是一种方法，公司能够缓冲自己从复杂多变的环境，同时仍然能够满足客户的需求。我们探讨了半导体行业在汽车行业的可变性。在供应方面，我们评估了制造过程复杂性的可变性，并且产品由多个部件组成，以随机生产提前期。然而，在本文中，我们忽略了供应方产生的可变性，因此我们假设交货时间固定为16周。对于需求端，这种现象被称为牛鞭效应，随着供应链的上升，需求的可变性增加，严重影响半导体供应链。这导致随机需求过程不能很好地理解。因此，我们从两个方面来评价需求的随机性：1)历史需求数据与其均值的离散度，表示为需求的标准差；2)实际需求与预测数据之间的差异，表示为预测误差的标准差。我们用它们作为需求变化的代理。然后我们将这些数据应用到基础股票模型中。然后，我们确定每个可变性参数对库存的贡献。库存模型用半导体公司提供的实际统计数据来表示半导体制造商的库存，以计算满足期望客户服务水平所需的库存目标。
As an innovation in the semiconductor industry grows speedy, supply chain processes have not followed up. The variability in semiconductor supply chain have increased and been more complicated. These results in accurately forecast demand and set inventory target. Demand and supply are more and more stochastic and non-stationary. Inventory is one of the methods that companies are able to buffer themselves from complex and variable environment, while still being able to satisfy customer needs. We explore the variability of semiconductor industry in automotive industry. On the supply side, we evaluate variability in complexities of manufacturing process and also products are composed with multiple parts efforts to stochastic production lead-time. However in this paper, we disregard the variability arising from supply side so we assumed lead-time is fixed at 16 weeks. For demand side, the phenomenon is known as the bullwhip effect, the demand variability increases as one move up a supply chain, severely effects to semiconductor supply chain. This results the stochastic demand process is not well understood. Thus we evaluate the stochastic in demand as two aspects: 1) the dispersion of historical demand data from its mean which denoted as standard deviation of demand, 2) the difference between the actual demand and forecast data which denoted as standard deviation of forecast error. We use them as a proxy for demand variability. Then we apply the data to the base stock model. Then, we determine what each variability parameter contributes to inventory. The inventory model represents the semiconductor manufactory’s inventory with actual statistical data which provided from semiconductor company to calculate inventory target required to meet the desired customer service level.