PURPOSE: To provide a continuous logic calculation system usable in the environment of wide ranges such as voice recognition, process control, optical character recognition, signal processing, image processing, communication and electronic equipment for domestic use.
CONSTITUTION: The continuous logic system using a neural network defines input and output variables (130) not to use any membership function, uses a generation rule (IF THEN rule) (132) for relating the output variables to the input variables and calculates or interpolates an output (136) by using the neural network. The neural network learns the prescribed generation rule, at first, and next generates the output in real time. The network is composed of artificial neurons and each artificial neuron has only one processing element in the form of a multiplier. The network utilizes an algorithm for applying an inclusive minimum value to a prescribed input vector set without requiring any repeated training.