Markov动态决策过程在耐用品购买中的应用  

Markov Dynamic Decision Model for Purchase Decision of Durable Goods

作　　者：贾俊秀[1] 

机构地区：[1]西安电子科技大学理学院,经济管理学院,西安710071

出　　处：《工程数学学报》2009年第5期827-835,共9页Chinese Journal of Engineering Mathematics

基　　金：国家自然科学基金(70701028)

摘　　要：通过建立一个离散型Markov决策模型,并在模型中引入感知价值和感知价值系数,本文研究了消费型耐用品购买的多阶段最优策略。当每个产品阶段中有两代不同产品时,消费者最优购买策略存在;在一定的条件下,从某个状态开始最优策略为"购买"。最后以一个数值实例验证了这些结论,并说明给定相关数据后可获得购买耐用品的最优决策,这一决策与消费者的偏好和卖方的定价存在密切关系。A discrete time Markov decision process model was built,introduced in the model are the concepts of perceived value and perceived coeffcient. The optimal policies to purchase consumptive durable goods were studied under multi-stages. When there are two generations in each product stage,the consumer s optimal purchasing policy exists. It is proved that exists some state,from which the consumer s optimal policy is ＂to buy the product＂ under some conditions. A numerical example show the reliability of the method.Given related data, the optimal purchase policies can be obtained, which have closed relation with the consumer＇s preference and the pricing of the bargainor.

关 键 词：MARKOV决策过程 耐用消费品 最优购买决策 感知价值 

分 类 号：O211.62[理学—概率论与数理统计] F224[理学—数学]