非平稳随机需求供应链自适应库存控制研究

发布时间：2018-01-14 07:36

本文关键词：非平稳随机需求供应链自适应库存控制研究　出处：《青岛大学》2015年硕士论文　论文类型：学位论文

【摘要】：供应链是由供应商、制造商、批发商、零售商及最终客户等相互联系的成员组成的网络系统。企业实施供应链管理的目的是实现企业横向一体化,通过企业协作,提高响应市场需求的速度。库存在供应链管理中扮演主要角色,直接影响整个供应链系统的成本和服务质量。供应链环境下的库存控制不但与单个企业本身密切相关,而且与供应链上的所有企业都密切相关的问题,需要以整体观念协同管理,以达到系统整体最优的目标。自适应库存控制就是以自适应控制理论的思想实现不确定环境下的库存控制问题。具体控制策略是:针对多级多周期库存控制问题,在客户需求信息不确定的情况下,通过库存盘点反馈库存控制效果,并根据库存效果与库存控制目标(如顾客服务水平)的差距及时调整控制方式,达到预设的库存控制目标。本文的研究目标是探讨非平稳需求环境下多周期自适应库存控制方法。主要研究内容包括:针对非平稳需求多周期库存控制问题,在信息共享和不共享条件下分别建立了两个自适应库存控制模型,基于强化学习算法设计库存控制策略,确定合理的订货时间和订货量,使得服务水平能够尽可能满足预先设定的目标,而且有效减轻牛鞭效应的不良影响。针对需求信息和服务水平评价方面的模糊性特征,在信息共享条件下,用模糊指数平滑法预测未来需求,用强化学习算法实现库存控制。最后分别通过计算机仿真验证设计的库存控制方法的有效性。仿真结果表明,所给出的控制方法可以使得服务水平到达较好的效果,同时,模糊指数平滑预测方法在减小牛鞭效应方面有较好的效果。
[Abstract]:The supply chain is composed of supplier, manufacturer, wholesaler, retailer and the end customer network system as mutual contact of the members of the enterprise. The implementation of supply chain management is to achieve the purpose of the enterprise through horizontal integration, enterprise collaboration, improve the response speed. The stock market demand plays a major role in supply chain management, directly affect the entire supply chain system the cost and quality of service. The inventory control under the supply chain environment not only with a single enterprise is closely related to itself, and all enterprises in the supply chain are closely related to the problem, the need to read the collaborative management in the view of the whole, to achieve the best overall goal. The adaptive inventory control is thought to the adaptive control theory to realize the control problem under the environment of uncertain inventory. The specific control strategy is: to solve the control problem for multi period inventory, the customer demand information uncertainty. Under the condition of the inventory, inventory control and feedback effect, according to the effect of inventory and inventory control target (such as customer service level) the gap adjust control mode, to achieve the inventory control target of presupposition. The goal of this study is to investigate the nonstationary demand under the environment of multi period adaptive inventory control method. The main research contents include: for nonstationary demand multi period inventory control problems in information sharing and sharing conditions were established two adaptive inventory control model, reinforcement learning algorithm design of inventory control strategy based on the determined order time and order quantity reasonable, the service level can be as much as possible to meet the target, but also reduce the adverse effects of bullwhip effect the fuzzy feature evaluation. According to the demand information and service level, in the condition of information sharing, prediction using fuzzy exponential smoothing method The future needs, using reinforcement learning algorithm to achieve effective inventory control. Finally, through the design of the inventory control method of computer simulation. The simulation results show that the control method can make the service level to reach good results, at the same time, the fuzzy exponential smoothing forecasting method has better effect in reducing the bullwhip effect.

【学位授予单位】：青岛大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F274

【共引文献】