摘要
在总结油价预测模型的基础上,视原油价格动态过程为马尔科夫链,利用数理统计方法计算得到马尔科夫链状态转移概率;根据马尔科夫链的性质,在其状态转移概率基础上,得到其长游程的稳态概率;继之利用动态规划技术,进行原油进口时机决策。利用这种方法对实际油价进行原油进口时机决策。结果表明,新决策方法能够有效规避'买涨不买跌',且简单易行,有效提高数据利用率。
By summarizing oil prediction models,the oil price dynamic progress was regarded as markov chain. The states transition probability of markov chain was calculated by the mathematical statistics method. According to markov-chain's property,the long-runs steady probability was obtained on the basis of state transition probability. The dynamic programming techniques was used to get opportune decisions of oil importing. The practical opportune decisions of oil importing can be done by this method. The results show that this decision method could avoid " to buy price rising instead of dropping" ,be simple and convenient,and improve the data use factor.
出处
《中国石油大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2010年第5期183-187,192,共6页
Journal of China University of Petroleum(Edition of Natural Science)
基金
中央高校基本科研业务费专项资金资助项目(R090629B)
关键词
原油价格
马尔科夫链
稳态概率
动态规划
进口策略
oil price
markov chain
steady probability
dynamic programming
import policy