摘要： 在许多情况下,知识源不可能给出确定的假设,它仅能给出假设的信任值(Belief Value)和不信任值(Disbc-lief Value).我们讨论了一种处理这种不确定性的方法,不信任值在这种方法中起着重要的作用,各个专家互相影响,通过改变自己对假设的可信度值(Certainty Facfor Value)CFV来进行合作.一个假设可信度值CFV的最终值是根据各可信度值的平均值和模糊熵来确定的.

关键词: 分布式知识库系统, 不确定性处理, 平均值, 模糊熵

Abstract: In many cases, KS (Knowledge Source) may not give definite hypothesis other than belief and disbe-lief value for hypothesis. A method of processing this uncertainty is discussed, in which disbelief value plays an im-portant part. Experts influnce each other and cooperate by modifying their CFV (Certainty Factor Value) for hy-pothesis. The final CFV for hypothesis is given based on mean and fuzzy entropy of all CFV.