Fuzzy reasoning with a rete-oo rule engine

Abstract

Rules and rule engines play an important role in automated decision making processes like business workflows or system monitoring. Classical inference machines evaluate rules until a final "yes" or "no" decision: this crisp classification schema can turn into a deficiency when they have to deal with uncertain or inprecise knowledge. To circumvent some of these limitations we have built the "Java Expert Fuzzy Inference System" (Jefis) and implemented factory methods to deploy the Jefis library as an extension for the classical rule engine JBoss Drools. We outline the new features and give examples of uncertain formulated rules executing within the Jefis Drools extender. © 2009 Springer-Verlag Berlin Heidelberg.