基于贝叶斯更新的供应链协同预测模型研究  

Study on Supply Chain Collaborative Forecasting Game Based on Bayesian Updating

作　　者：魏炜[1] 申金升[1] 

机构地区：[1]北京交通大学交通运输学院,北京100044

出　　处：《预测》2010年第5期68-73,共6页Forecasting

基　　金：国家科技攻关计划资助项目(2004BA205A29)

摘　　要：本文运用纳什均衡和贝叶斯更新模型,得到了供应链联合预测均衡的存在条件。模型中,供应商和零售商均需决定是否投资于预测和信息分享技术,双方的需求预测将由零售商汇总成一个统一的预测。结果表明,当考虑信息分享成本时,供应商在投资后必然会分享,而零售商不会进行信息分享。当双方预测之间相关性较低、双方预测能力均处于中等水平、双方谈判实力比较接近时,实现联合预测的可能性较大。这一结论为企业是否参与CPFR项目以及选择协同预测伙伴提供了决策依据。This paper adopts Nash equilibrium and Bayesian updating model,obtains conditions that favor supply chain collaborative forecasting.In the model,supplier（S）and retailer（R）should decide whether to invest in forecasting and information sharing technology.Their forecasts will be combined to form a single forecast by R.The results show,when taking information sharing cost into account,S always share after invest,while R never share.Joint forecasting is more likely to realize when the similarity between both partners＇ signals is low,forecasting capability is at intermediate level,and both partners have similar negotiation power.The results provide foundations for companies to decide whether to participate in CPFR initiatives and whether to choose certain partner to forecast collaboratively.

关 键 词：贝叶斯更新 信息分享 博弈论 协同预测 CPFR 

分 类 号：F272.1[经济管理—企业管理]