【作者】 王颖；

【导师】 王丽亚； 陶蕾芳；

【作者基本信息】 上海交通大学 ， 物流工程（专业学位）， 2017， 硕士

【摘要】 在商品差异化日益缩小的今天,市场的竞争越来越多地转变为供应链的竞争。对于工业机械制造商而言,这种供应链的竞争更加直接地反映在售后服务的竞争上。而在售后服务中,所需的零备件有及时的库存响应是首要前提。本课题来源于居于全球前列的空气压缩机技术企业A公司亚太区物流分拨中心关于如何有效地准备库存以应对市场需求的思考。目标是使预备的库存水平契合市场需求的波动,同时保有优良的库存周转率。本文基于实际情况,把涉及主营业务的43267项物料纳入研究范围(其中10527项在过去的一年中有历史销量)。通过两次运用ABC分类法,以对库存满足率和库存成本的影响为依据,进行物料分类,确立重点物料。随后,本文根据历史需求类型对重点物料进行细分,选取波动型消耗类别和快速型消耗类别中的两个典型物料进行预测方法研究。在对波动型消耗类的典型物料的研究中,本文针对其带有显著的季节性特征,用时间序列法的乘法模型对其进行预测,并用滚动计算季节指数的方法进行预测的修正,提高误差精度;在对快速型消耗类的典型物料的研究中,本文提出将移动平均法与指数平滑法相结合有助于获取更准确的预测结果。最后,本文尝试针对重点物料的预测方法进行普适性的推广,总结月度库存水平调整工作的方向和思路。本课题的意义不仅在于确立重点物料和预测方法,保持有效库存水平,还在于为集团尚在开发中的库存水平自动调整工具提供参考思路,减轻月度库存水平管理中的人工判断的工作量,降低主观因素的影响。更多还原

【Abstract】 As differentiation among products getting smaller and smaller among competitors in the market nowadays,the competition is more and more on the response of supply chain.For the market of industrial machines,the severe challenge from the requirement of response acts much on the after-sales service,which brings high expectation on the availability of stock.The study of this paper initiated by company A,who sits as international leader in compressor manufacturing,aims for having good stock level management in the distribution center of Asia Pacific in order to better serve the after-sales market.Based on the real situation,we take 43267 items related to core business into this study.(10527 items of them were with sales history in the past 12 months).By using 2 times of Activity Based Classification method,taking the impact on availability and inventory cost into consideration,we define the key group of items,from which,based on the historical demand pattern we pick up 1 typical item from erratic movers and 1 typical item from fast movers,to discuss the proper forecasting methods.For the erratic example,since it shows obvious seasonal feature,we apply time series recount multiplicity model to the forecasting,by calculating rolling seasonal indexes,the accuracy of forecasting is improved;for the fast mover,the study proposes to combine moving average method with exponential smoothing to get a better forecast result.At the end,we try to promote the methods to the key group upon which we summarize the way of working for stock level management.Not only this study contributes on the way of forecast and an efficient stock level,but also as the reference idea on developing a system tool for stock level management,which will help reduce the workload of stock level adjustment and the impact of human factor.更多还原