文章编号:1000-1964(2005)02-0261-04
股票投资的马尔可夫决策规划模型
韩 苗,薛秀谦,周圣武,康建林
(中国矿业大学理学院,江苏 徐州 221008)
摘要:应用马尔可夫决策规划理论,讨论了一种股票动态投资策略,将股票价格随机时间序列分解成趋势序列和残差序列两部分之和.在验证残差序列具有马尔可夫性的基础上,对其建立模型并进行投资决策.所给定理保证了在一定条件下该模型目标函数最优投资策略的存在,同时给出了求解最优策略的算法,并进行了二阶段算例分析.最后,通过具体实例验证了该投资决策模型的可行性.
关键词:马尔可夫决策规划;x2检验;最优策略;股票
中图分类号:O 29;F 830.91 文献标识码:A
Markov Decision Programming Model for Stock Investment
HAN Miao, XUE Xiu-qian, ZHOU Sheng-wu, KANG Jian-lin (School of Sciences, China University of Mining & Technology, Xuzhou, Jiangsu 221008, China)Abstract: By applying the Markov decision programming theory, a dynamic investment strategyfor stocks was discussed. The stochastic time series of stock price was decomposed into the sum oftendency series and residual series. And it was proved that the latter has a Markov property, as aresult, a model for investment decision was established. The theorem proposed showed that theoptimal investment strategy of the target function exists under certain condition. In addition, analgorithm to seek optimal strategy and an example of two stages were given. The feasibility of thisinvestment decision model was proved by some practical examples.
Key words: Markov decision programming; x2 test; optimal strategy ; stoc
k
文章编号:1000-1964(2005)02-0261-04
股票投资的马尔可夫决策规划模型
韩 苗,薛秀谦,周圣武,康建林
(中国矿业大学理学院,江苏 徐州 221008)
摘要:应用马尔可夫决策规划理论,讨论了一种股票动态投资策略,将股票价格随机时间序列分解成趋势序列和残差序列两部分之和.在验证残差序列具有马尔可夫性的基础上,对其建立模型并进行投资决策.所给定理保证了在一定条件下该模型目标函数最优投资策略的存在,同时给出了求解最优策略的算法,并进行了二阶段算例分析.最后,通过具体实例验证了该投资决策模型的可行性.
关键词:马尔可夫决策规划;x2检验;最优策略;股票
中图分类号:O 29;F 830.91 文献标识码:A
Markov Decision Programming Model for Stock Investment
HAN Miao, XUE Xiu-qian, ZHOU Sheng-wu, KANG Jian-lin (School of Sciences, China University of Mining & Technology, Xuzhou, Jiangsu 221008, China)Abstract: By applying the Markov decision programming theory, a dynamic investment strategyfor stocks was discussed. The stochastic time series of stock price was decomposed into the sum oftendency series and residual series. And it was proved that the latter has a Markov property, as aresult, a model for investment decision was established. The theorem proposed showed that theoptimal investment strategy of the target function exists under certain condition. In addition, analgorithm to seek optimal strategy and an example of two stages were given. The feasibility of thisinvestment decision model was proved by some practical examples.
Key words: Markov decision programming; x2 test; optimal strategy ; stoc
k