一种基于任务分解的多知识库协同求解专家系统  

A Rule-Based Expert System with Multiple Knowledge Databases

作　　者：宋良图[1] 刘现平[1] 毕金元[1] 查金水[1] 

机构地区：[1]中国科学院合肥智能机械研究所

出　　处：《模式识别与人工智能》2006年第4期515-519,共5页Pattern Recognition and Artificial Intelligence

基　　金：国家863计划重点资助项目(No.2003AA118070)

摘　　要：针对特定领域的知识特点、知识表示方法及采用的推理模型,提出一种基于产生式规则的多知识库专家系统。该系统改进传统专家系统的框架设计,根据求解问题的类别划分将知识库分解成相应的子知识库,再将子知识库的知识规则按知识表示的深度加以分解,建立反映专家经验知识的浅层知识库和原理性知识的深层知识库。系统采用主推理机和从推理机二级推理方式,不同的子知识库采用相应的从推理机。从而任务单一,搜索范围减小,能快速形成待检目标集。主从推理机制与正反向推理结合,提高系统的推理效率。运用该系统模型建造的农业领域专家系统实例,运行效率得到改善,速度显著提高。In this paper, a new rule-based expert system with multiple knowledge databases is designed according to the characteristics of the domain expert knowlege and its representation. Compared with classic expert system , the new system has a number of corresponding knowledge bases according to the sub-problems of a complex problem. And each knowledge base includes a shallow knowledge base of the empirical knowledge and a deep knowledge base of the scientific knowledge. The inference engine of the system adopts two-level inferring structure, the inference at system level and the one of sub-problem, which can reduce the searching space quickly and work efficiently. The system runs well over internet.

关 键 词：专家系统 多知识库 任务分解 主从推理 框架设计 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]