论文部分内容阅读
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。
Aiming at the knowledge characteristics, knowledge representation methods and reasoning models in a particular field, a multi-knowledge base expert system based on production rules is proposed. The system improves the traditional expert system framework design, decomposes the knowledge base into corresponding sub-knowledge bases according to the category of solving problems, and then breaks down the knowledge rules of sub-knowledge bases according to the depth of knowledge representation, Layers Knowledge base and the knowledge base of the deep knowledge base. The system uses the main reasoning machine and reasoning machine two reasoning methods, different sub-knowledge base using the corresponding from the inference machine. Thus a single task, the search area decreases, can quickly form the target set to be seized. Master-from-reasoning mechanism and positive and negative reasoning combine to improve the system’s reasoning efficiency. The application of this system model to agricultural expert system has improved the operational efficiency and significantly increased the speed.