混合案例推理和规则推理的集成推理框架

Hybrid reasoning framework integrating of rule-based and case-based reasoning

在智能决策系统(IDSS)获取知识的推理体系中,案例推理和规则推理有着各自的优点,而混合两者的集成推理可以克服两者的缺点,提高系统的效率和综合推理能力。但是集成推理系统缺乏通用性,延长了开发周期,且不利于规则库和案例库的重用。一种可扩充的集成推理框架为了解决上面的问题而被提出,该框架利用智能决策支持语言Knonit的组件性,对不同的集成方式可方便地扩充相应的集成推理方案,从而快速地搭建IDSS应用;同时规则和案例是作为Knonit广义知识元存在,可以在集成推理框架中复用,另一方面,Knonit的动态特性和可扩充性也对案例库和知识库动态的调整和扩充提供了支持。

Hybrid reasoning system integrating of rule-base reasoning and case-base reasoning will greatly improve the efficiency oflDSS system and the general reasoning ability. Because of the incompatible of hybrid system, the IDSS developing cycle is postponed. On base of the IDSS environment language Knonit, an extensible hybrid reasoning framework was introduced, With the component characteristic of Knonit, reasoning solution could be added to cope with the new integrating way, and the rule library and case library could be highly reused.