Mathematical model of the neural network conclusion defuzzificator in fuzzy-logic output procedures and its software implementation

This paper presents a mathematical model of a neural network defuzzificator. It is a twolayer perceptron and serves to convert a fuzzy solution to a numerical form in fuzzy logic derivation procedures. The model allows to optimize the computational load that occurs when using the standard center of gravity method, through the use of a neural network. Training and testing was conducted with various settings of the neural network model. The effectiveness of this approach with measuring the time of computing operations was also proved.