不确定环境下多周期VMI自适应库存控制方法  

An Adaptive Inventory Control for a Multi-Period VMI Model with Non-Stationary Demand

作　　者：孙超[1] 龚卓[1] 张德胜[1] 张纪会[1] 

机构地区：[1]青岛大学自动化工程学院复杂性科学研究所,山东青岛266071

出　　处：《青岛大学学报（工程技术版）》2011年第3期32-36,共5页Journal of Qingdao University(Engineering & Technology Edition)

基　　金：山东省自然科学基金资助项目(ZR2010GM006)

摘　　要：随着供应链中不确定性因素的增加,供应商管理库存(vendor managed inventory,VMI)作为一种新型的库存管理方式,正被越来越多的企业和学者所关注。针对单供应商单零售商的VMI模型,在顾客需求未知的情况下,提出了一种利用二次指数平滑法预测需求,用强化学习算法评估安全因子的效果,进而调整顾客服务水平的自适应库存控制方法。仿真结果表明,在非平稳需求情况下,采用该自适应库存控制策略,能够得到比较满意的顾客服务水平。With more and more uncertainties appeared in supply chains,the VMI inventory management as a new inventory control approach was concerned by more and more businesses and academicians.In this paper,a VMI inventory control model for a two-stage supply chain consisted of one supplier and one retailer was proposed.For non-stationary demand,an adaptive inventory control method was proposed.In this approach,the secondary exponential smoothing method was used to forecast the demand,and the reinforcement learning was adopted to evaluates the effect of the adopted safety factor,and then to adjust the level of customer service.Simulation results show that the proposed adaptive inventory control scheme can get a very satisfied service level for different non-stationary demand mode.

关 键 词：供应商管理库存 顾客服务水平 强化学习 安全因子 

分 类 号：F253.4[经济管理—国民经济] F224