推理机制是专家系统的核心,系统的控制效率是系统成功的一个重要的方面,其算法选择的合适与否直接影响整个系统的执行效率。本系统采用里德模式匹配算法(Rete PatternMatching Algorithm)。Rete算法的基本思想是保存过去匹配循环中留下的全部信息,以空间代价来换取产生式系统的执行效率。对每一个模式,附加一个匹配事实表来记录动态数据库中能与它匹配的事实。当一个新的事实加入到匹配事实表中,算法能找出所有能与之匹配的模式,并将该事实加入到匹配事实表中去;当一个事实从动态数据库中删除时,同样找出所 Inference mechanism is the core of expert system. The efficiency of system control is an important aspect of system success. Whether its algorithm selection is appropriate or not will directly affect the efficiency of the whole system. The system uses Reed Pattern Matching Algorithm (Rete PatternMatching Algorithm). The basic idea of ​​Rete algorithm is to save all the information left in the past matching cycle, in exchange for the space cost in exchange for the implementation efficiency of the production system. For each pattern, attach a matching fact table to record the fact that it matches with it in the dynamic database. When a new fact is added to the matching fact table, the algorithm can find out all the patterns that match with it and add the fact to the matching fact table. When a fact is deleted from the dynamic database,