Filtration of non-monotonic rules for fuzzy rule base compression

This paper proposes a rule base compression method for fuzzy systems. The method is based on filtration of rules with identical linguistic values for the output that are known as non-monotonic rules. The filtration removes the redundant computations in the fuzzy inference with respect to the crisp values of the inputs to the fuzzy system. The method identifies the redundant rules after fuzzification and removes them while preserving the defuzzified output from the fuzzy system for each simulation cycle. In comparison to the known rule base reduction methods, this rule base compression method does not compromise the solution and has better efficiency in terms of on-line computations. The method processes the rule base for a fuzzy system during simulation cycles by contracting it to a rule base of a smaller size at the start of each inference stage and then expanding it to its original size before the next fuzzification stage.