基于需求预测的电商供应网络库房选址和库存策略
发布时间：2018-03-16 10:16
本文选题：电商供应网络 切入点：需求预测 出处：《北京工业大学》2016年硕士论文 论文类型：学位论文
【摘要】：互联网时代,电子商务发展越来越快,在电商平台上,如京东,亚马逊,自营产品种类繁多,采购、交付活动频发,与上游供应商和下游用户组成的供应网络越来越复杂。如何管理好电商供应网络中的库存,以期更大程度满足用户需求已成为电商供应网络发展的目标。电商供应网络的各个库房除了要满足服务区域的用户需求,还需要支持其他库房,通过库存合作转运来降低供应网络的库存水平,提高用户满意度。文章主要针对电商供应网络中快速消费类产品展开研究。对比分析市场需求预测方法的优劣势,选择科学的需求预测方法。基于需求预测,考虑随机需求和库存转运机制,通过库存配置和转运研究电商供应网络的库存策略和转运策略。首先,从不同视角下对市场需求进行分析,归纳总结宏观、中观、微观层次市场需求的特点,对比不同层次市场需求下几种预测方法的优劣势。研究发现考虑购买量,点击量,库存量等因素,并根据节假日、交通、天气等特殊情景进行分析调整构建的综合集成预测模型的效果最好,更能够准确预测宏观和中观市场需求,而BP神经网络在预测波动性较小的微观市场需求方面具有优势。其次,针对电商供应网络的支持城市,通过改进的K-means聚类和多属性决策理论,基于最快满足用户需求准则,确定最大降低库存成本的库房选址方案。并对比分析本文所提出的选址方案和A公司目前选址方案的差异性,对比库存成本发现提出方案在理论上能够能降低成本,提高用户服务满意度。再次,根据市场需求预测,基于存储成本和缺货损失风险,得到库房的最优基本库存策略,最大化企业利润。同时,研究发现多库房多产品情形下,当发生缺货时,通过库存转运策略可以协调供应网络库存水平,降低库存成本,采用库存转运策略比不采用更能提高用户服务和供应网络利润。结合A公司的实际数据进行实证分析,结果表明新的库存选址方案在满足用户交货服务的基础上,降低了企业成本,同时缩短交货周期,提高用户服务体验。通过库存转运机制,能更好的实现库存均衡,减少库存成本,提高用户满意度。
[Abstract]:The age of the Internet, electronic commerce develops more and more quickly, in the electronic business platform, such as Jingdong, Amazon, proprietary product variety, procurement, delivery and supply network activities frequently, with upstream suppliers and downstream users more complex. How to manage the electricity supply network inventory, to a greater extent in order to meet the needs of users electricity supply network has become the goal of development. Each warehouse supply network business in addition to meet the needs of users of the service area, also needs the support of other warehouse, to reduce the inventory level of supply network through inventory cooperation transshipment, improve customer satisfaction. This article mainly aims at fast consumer products business in the supply network is researched. A comparative analysis prediction method the market demand of the advantages and disadvantages, choose scientific methods of demand forecast. Based on demand forecasting, considering stochastic demand and inventory transport mechanism through the inventory coordination The transport of electricity supplier supply network inventory strategy and transport strategy. First of all, from the analysis of different perspectives on market demand, summed up the macro, meso and micro level market demand characteristics, several prediction methods of comparative advantages and disadvantages of different levels of market. The study found that the amount of clicks, consider purchasing, inventory factors so, according to the holidays, traffic, weather and other special situations for analysis of the integrated prediction model of the adjustment effect is the best, a more accurate prediction of macro and meso market demand, and the BP neural network has advantages in the micro market demand forecast volatility smaller. Secondly, according to the electricity supply network to support the city, by K-means improved clustering and multi attribute decision making theory, based on the criterion to determine the maximum to meet the needs of users, reduce the cost of inventory warehouse location scheme. And comparative analysis 鏈枃鎵，
本文编号：1619478
本文链接：https://www.wllwen.com/jingjilunwen/guojimaoyilunwen/1619478.html
