Please choose your delivery country and your customer group
A production system is often used as the basic architecture in artificial intelligence or an expert system. In general production systems, a pattern matching process consumes a large part of the time required for reasoning. To improve this, the Rete algorithm has previously been proposed. The Rete algorithm uses knowledge concerning the similarity among productions and the degree of the production's satisfaction with the contents of work-memory, but it has a considerable wasteful matching process. Therefore, the authors propose an enhanced pattern matching algorithm which adds knowledge concerning the exclusiveness of each production and the matching success rate. Finally, they prove the effectivity of their proposal by showing three sample programs running on a 16-bit personal computer.