Tuning membership functions, tuning AND and OR operations, tuningdefuzzification: which is best?

摘要

Fuzzy control is a methodology that transforms linguistic IF-THENrules formulated by an expert into a precise control strategy. There aremany fuzzy control models that are able to implement suchtransformations. For a given situation, one model may perform betterthan another; as the state changes, switching between models can improvethe quality of the resulting control. However, the switching algorithmcan be computationally expensive. In this paper, we show that a similarincrease in the quality of the resulting control can be obtained (with amuch smaller computation time) by dynamically changing defuzzificationexclusively