一种HAL匹配算法在专家系统中的应用研究

Application of HAL Matching Algorithm in Expert System

提出一种新的HAL(Heuristically-Annotated-Linkage)匹配算法来解决传统的RETE和TREAT算法在专家系统中匹配网络过多、匹配时间较长的问题。HAL通过建立相关规则和类的启发式反馈通道,减少了冗余内容和匹配网络。与以上算法运用规则的启发式信息不同的是HAL主要使用类的启发式信息,因为在一个程序中,规则的数量要比类多,类的数量一般是固定的。HAL只需建立一个全局匹配网络,而RETE和TREAT算法要对所有规则建立相应的匹配网络,所以,它解决了应用RETE和TREAT算法时匹配网络过多,匹配时间较长的问题。

A new HAL （Heuristically-Annotated-Linkage） matching algorithm is proposed to solve the problems of the traditional RETE and TREAT algorithm in the expert system to match the network with a long time, the matching time is longer. HAL reduces the redundancy content and the matching network by establishing the relevant rules and class of heuristic feedback channels. Different from the heuristic information of the rules of the above algorithm, the main use of the heuristic information is HAL, because in a program, the number of rules is more than the number of classes, the number of classes is generally fixed. HAL only needs to establish a global matching network, and the RETE and TREAT algorithm to establish the corresponding network, so it solves the problem of RETE and TREAT algorithm, the matching network is too much and the matching time is longer.