摘要
在进一步发展认知图、影响图理论的基础上，提出了融合二者于一体的认知影响图理论，建立了两种构造认知影响图方法：诊断法和时间序列法，解决了实际认知过程中，按时间序列出现事件经常与人类诊断知识方向矛盾的问题：并根据在思考问题过程中，同一领域专家的认知方式不同，分别采取因果推理及诊断推理的特点。
In this dissertation, the theory of cognitive influence diagrams, which is constructed by integrating cognitive maps and influence diagrams, is proposed. Two new methods called time series support and diagnostic support are originally presented to structure cognitive influence diagrams. These methods can solve the problem that chronology in which events occur during the actual execution of a plan often clashes with the directionality of diagnostic relationships along which human knowledge is organized. In addition, we notice the cognitive differences in the thought processes of the different experts. A new propagation mechanism that can acquire causal and diagnostic supports is presented.
出处
《系统工程学报》
CSCD
1997年第1期87-95,共9页
Journal of Systems Engineering
基金
国家自然科学基金
关键词
认知影响图
概率衍值
智能推理
人工智能
cognitive influence diagrams, probability propagation, intelligent reasoning
