摘要
运用证据理论研究不确定性的表示问题.首先给出了证据获取过程的形式描述,区分了命题的不确定性和证据的不确定性.继而通过引入证据理论的基本概念,指出了构成证据不确定性的三大因素:不完全,不精确和不肯定.随后构造了一个描述不确定性的指标函数U,它是从三维不确定性空间到[0,1]区间的一个映射,最后进行了若干分析讨论,提出并证明了证据的合成单调性.
The representation of uncertainty is studied using the evidence theory. Some fundamental concepts and the combining rule in the evidence theory are introduced. A formalized description of the evidence acquisition process is given. Proposition uncertainty and evidence uncertainty are distinguished. Three factors that constitute the evidence uncertainty and reflect the uncertainty in actual problems are pointed out. A new method of describing the evidence uncertainty, the index function method, is discussed. According to given properties of the index function, such a function has been developed for three different cases and a unified index function is obtained. As a measure of evidence uncertainty, the index function U is a mapping from three-dimensional uncertainty space onto interval [0,1]. Discussions are made and the monotonicity for evidence combination is proved. The curve Umax is given and an extension.of the index function to a weighted one is proposed.
出处
《华中理工大学学报》
CSCD
北大核心
1993年第3期24-30,共7页
Journal of Huazhong University of Science and Technology
基金
国家"863"高技术计划资助项目
关键词
人工智能
证据理论
不确定性
artificial intelligence
evidence theory
evidence uncertainty
combination monotonicity
index function method
uncertainty analysis
information processing
