摘要
本文首先分析了传统定性推理方法中存在的歧义性分枝和建模手段不强的问题。在此基础上,介绍了定性推理与不完全定量知识相结合的方法,包括它与区间分析、模糊逻辑、数值模拟和量级思想相结合的方法,这些结合方法皆有助于解决上述的两个问题。最后,本文还指出定性推理的应用领域已从最初的物理领域迅速扩展到医学、经济和商业等领域。
In this paper, the problem of ambiguous branches and the weakness of modeling measures existed in the classical qualitative reasoning are analyzed at first. Based on these, we introduce the methods which combine the qualitative reasoning with interval analysis, fuzzy logic, numerical simulation technique and reasoning of order of magnitude respectively, all these methods can help to solve the two problems mentioned above. Finally, we emphasize that the application areas of qualitative reasoning have been extended from the physics to the medical science, economics and business.
关键词
人工智能
定性推理
定性模拟
定性代数
模糊逻辑
artificial intelligence, qualitative reasoning, qualitative simulation, qualitative algebra, fuzzy logic