| Retrospect and Prospect of the Influence Diagram |
| |
| 作者姓名： | Liu Yanqiong Shen Yongping Chen Yingwu |
| |
| 作者单位： | SchoolofInformaionSystemandManagement,NationalUniversityofDefeneeandTechnology,HunanChangsha410073 |
| |
| 基金项目： | NationalScienceFoundationofChina (No .70 2 72 0 0 2 ) |
| |
| 摘 要： | The evaluation algorithm and the application of the influence diagram were surveyed, which argues that to construct an explicit,compact and objective influence diagram is of the most importance. There are two suggested ways for realization of the influence diagram: introducing the achievements of the modern psychology, cognitive science, behavior science, and so on to represent and solve uncertainty to build a well-constructed influence diagram; based on the observed data to build an influence diagram. Also, the limitations of the influence diagram were analyzed, such as that it cannot deal with asynunetric problems efficiently, cannot picture dynamic problems,cannot model the problems with a limitless horizon, and ther is no highly efficient algorithm. And some potential methods to overcome these limitations were pointed out.
|
| 关 键 词： | 影响图 限度 应用 DAG 心理学 认知科学 行为科学 |
Retrospect and Prospect of the Influence Diagram |
| |
| Liu Yanqiong,Shen Yongping,Chen Yingwu.Retrospect and Prospect of the Influence Diagram[J].Engineering Sciences,2005,3(1):58-63. |
| |
| Authors: | Liu Yanqiong Shen Yongping Chen Yingwu |
| |
| Institution: | (School of Informaion System and Management, National University of Defence and Technology, Hunan Changsha 410073) |
| |
| Abstract: | The evaluation algorithm and the application of the influence diagram were surveyed, which argues that to construct an explicit, compact and objective influence diagram is of the most importance. There are two suggested ways for realization of the influence diagram: introducing the achievements of the modern psychology, cognitive science, behavior science, and so on to represent and solve uncertainty to build a well-constructed influence diagram; based on the observed data to build an influence diagram. Also, the limitations of the influence diagram were analyzed, such as that it cannot deal with asymmetric problems efficiently, cannot picture dynamic problems, cannot model the problems with a limitless horizon, and ther is no highly efficient algorithm. And some potential methods to overcome these limitations were pointed out. |
| |
| Keywords: | influence diagram evaluation algorithm application limitation |
| 本文献已被 CNKI 维普 等数据库收录！ |
| 点击此处可从《工程科学》浏览原始摘要信息 |
| 点击此处可从《工程科学》下载免费的PDF全文 |
|
