The Performance Optimization of CLIPS

CLIPS is an expert system tool which provides a complete environment for the construction of rule and/or object based expert systems. Because of its portability, extensibility, capabilities, and low-cost, CLIPS has received widespread acceptance. As a production system CLIPS´s performance rapidly decreases as the number of working memory elements increases. To address this problem, this paper is aimed at speeding up CLIPS in the case of dealing with large amount of facts and changeable facts. In this paper two measures are adopted to improve CLIP´s performance, firstly RETE algorithm which is used as the rule condition-testing algorithm in CLIPS are replaced with TREAT algorithm. Secondly some practical techniques are employed when implementing TREAT algorithm such as recording partial matches, hashing the alpha node, adding rule group. Experimental results show that the two measures can effectively improve CLIPSpsilas performance.