| 引用本文: | 章辉, 孙优贤.离散随机模型降阶的最小条件信息损失方法[J].控制理论与应用,2005,22(6):919~924.[点击复制] |
| ZHANG Hui,SUN You-xian.Minimum conditional information loss approach to discrete stochastic model reduction[J].Control Theory and Technology,2005,22(6):919~924.[点击复制] |
|
| 离散随机模型降阶的最小条件信息损失方法 |
| Minimum conditional information loss approach to discrete stochastic model reduction |
| 摘要点击 1303 全文点击 2031 投稿时间：2003-10-30 修订日期：2005-05-13 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/j.issn.1000-8152.2005.6.013 |
| 2005,22(6):919-924 |
| 中文关键词 模型降阶 状态集聚 最小信息损失 熵 离散线性随机系统 |
| 英文关键词 model reduction state aggregation minimum information loss entropy discrete linear stochastic system |
| 基金项目 国家973计划资助项目(2002CB312200). |
|
| 中文摘要 |
| 针对随机系统的模型降阶问题,从分析离散线性随机状态方程模型中的条件信息描述机制入手,讨论了模型状态集聚过程中系统的平均条件信息损失.运用在模式识别领域中获得成功应用的最小信息损失准则得出了一种新的模型降阶信息论方法———基于状态集聚的最小条件信息损失方法,并讨论了降阶模型阶次的选择.分析表明,当原系统是渐近稳定时,由该方法得出的降阶模型也是渐近稳定的.该方法运用简单,仿真研究也表明由该方法得出的降阶模型具有良好的近似性能. |
| 英文摘要 |
| Aiming at the problem of stochastic model reduction,the average conditional information loss of the state aggregation is discussed by analyzing the conditional information description of discrete linear state space model.A new information theoretic method for linear stochastic model reduction,the minimum conditional information loss approach based on state aggregation,is deduced by the principle of minimum information loss which was applied successfully in the field of pattern recognition.The selection of the order of reduced-order model is also discussed.It is proved that the derived reduced-order model is asymptotically stable when the full-order model is asymptotically stable.Simulation results show that,this method is easily applicable and the reduced-order models derived by it have good approximation performance. |