中文摘要：

在农业病虫害诊断领域，传统的专家系统往往在设计时就硬编码好了知识结构，在不重新改动代码的情况下，具有很差的扩展性。介绍了一个基于本体和案例推理的可重构知识管理框架ReKM。该框架利用本体作为知识结构的描述，由知识存储层、本体层和应用层组成。利用本体模型的可重构特性，一定程度上克服了传统专家系统可扩展性不足的缺点。文中详细阐述了框架中基于案例的检索算法和简单推理的概念。最后介绍了基于该框架开发的一个系统在某农科院病虫害诊断领域的应用情况和今后研究的方向。

英文摘要：

In the domain of diagnosis of plant diseases and insect pests, traditional expert systems hardcode the knowledge structure during design period, so they are not easy to be extended without rewriting the source code. This paper introduces an ontology - and - case - based reconfigurable knowledge management framework - ReKM. The framework utilizes ontology as the description of knowledge structure and is composed of three layers, which are knowledge storage layer, ontology layer and application layer. Due to the reconfigurablity of the ontology model, it, to some extent, overcomes the lack of extensibility in traditional expert systems. Then the algorithm of case - based retrieving and concepts of simple reasoning are explained in detail. In the end, the situation where a system developed on the basis of this framework has been applied in the diagnosis of plant diseases and insect pests in an academy of agricultural sciences, and directions of future research, are introduced.