论文部分内容阅读
不确定性推理是人类思维活动中最本质的东西,同时也是人工智能所研究的重要内容。已有的不确定性表示、度量以及推理方法,由于推理模型本身不完备,所考虑的不确定性因素是不全面的,只能适用于一些简单场合的推理。本文在全面剖析了产生式规则中各类命题的不同含义之后,提出了通用不确定性匹配算法、通用不确定性更新算法,形成了完整的通用不确定性推理(GUR)模型。GUR模型是非常完善的不确定性推理模型,首次区分开了六类命题的不同含义,澄清了以前的模糊概念。在通用不确定性匹配算法和通用不确定性更新算法中,引入了几个新函数,使推理过程更加清晰,更加合理。同时GUR模型也为进一步研究具体的不确定性推理方法奠定了基础。
Uncertainty reasoning is the most essential thing in human thinking activities, and it is also an important part of artificial intelligence research. The existing uncertainties indicate that the uncertainties considered in the measurement and reasoning methods are not comprehensive and can only be applied to reasoning in simple cases because the inference model itself is not complete. After comprehensively analyzing the different meanings of the various propositions in the production rules, this paper proposes a general Uncertainty Matching (GUR) model, which is based on generalized Uncertainty Matching and General Uncertainty Updating algorithms. GUR model is a very perfect model of uncertainty inference, for the first time to distinguish the different meanings of the six types of propositions, to clarify the previous fuzzy concept. In the general uncertainty matching algorithm and general uncertainty updating algorithm, several new functions are introduced to make the reasoning process clearer and more reasonable. At the same time, the GUR model lays the foundation for further research on the specific method of uncertainty reasoning.