POAD专家系统中的知识检索方法

Knowledge retrieval methods for POAD expert system

为了充分发挥面向问题分析与决策(POAD)专家系统在大容量知识库构建和以人为主导推理方面的优势,需要增强系统的知识检索能力。根据POAD专家系统的知识表示和推理方法设计了简单检索及层次检索方法,简单检索方便,用户不需要特别训练;层次检索则可以实现复杂知识检索,对于大容量知识系统效果尤为明显。此外还设计了模糊检索和结果中再检索等其他检索方法。应用实验表明,在POAD专家系统中设计和开发有效的知识检索方法可以极大地提高专家系统的推理速度。

In order to bring the advantages of POAD expert system in construction of knowledge-base with large capacity and inference with the human-dominant manner into full play, the ability of knowledge retrieval of system needs to be enhanced. According to the knowledge representation and inference method of POAD expert system, this paper designed two knowledge retrieval methods, i. e. simple retrieval and hierarchical retrieval. Simple retrieval method was convenient and any special training for the user wasn＇ t needed, while hierarchical retrieval method could realize complex retrieval requirements and was often used to the large capacity knowledge system. Furthermore designed other retrieval methods, such as fuzzy retrieval and re-retrieval in results. Experimental results show that the knowledge retrieval methods designed and developed in POAD expert system can improve the inference speed greatly.