定性推理综述（Ⅰ）──常识知识推理和因果性解释以及传统定性推理方法

An Overview on Qualitative Reasoning (Ⅰ)──Commonsense Reasoning, Causal Explanation and Classical Qualitative Reasoning Methods

在人工智能的今天，知识已被认为是任何智能系统的集成元件。对常识知识推理和这种推理所需实践知识的研究已是并将在一段时间内仍会是人工智能研究的中心问题。本文在讨论常识知识推理和因果性解释两个概念的基础上，较详细地介绍了传统定性推理中三种主要方法：定性微分方程法、定性进程理论和定性模拟方法，并对它们的特点和共同之处予以总结。

Within the framework of the dominant paradigm in artificial intelligence today,knowledge is considered to be an integral component in any intelligent system. Studies of commonsense reasoning and the practical knowledge required for such reasoning have been and remain the central problem in artificial intelligence. In this paper, based on the concepts of commonsense reasoning and the causal explanation, we introduced the three main approaches of classical qualitative reasoning: A Qualitative Physics based on Confluence, Qualitative Process Theory and Qualitative Simulation in detail, and summarized their common characteristics in terms of the properties owned by each approach.