Minimization of fuzzy systems based on fuzzy inference graphs

In a large fuzzy rule-based system, a great deal of computation time is required for a fuzzy inference engine. A given fuzzy rule-based system is modeled as a fuzzy inference graph where each node in the graph corresponds to a relation representing a rule in the rule-based system. This paper presents algorithms to minimize the number of nodes in the graph using fuzzy operations as well as their properties to reduce the computation time of each inference. The algorithm sorts a graph into stages, iteratively applies the two major operations, fuzzy union as well as composition and results in the new graph with the minimum number of nodes without increasing the dimensionality of each node.