PURPOSE: A self-organization learning petri net is provided to obtain an optimal output value for an input value, and to realize the fast learning speed and the correct modeling. CONSTITUTION: A sample set having the already known input and output value is trained in order(S310). After setting up the first system parameter through the training of the first sample(S320), the second sample is trained by the system organized by the first system parameter(S330). A critical value is compared by comparing an error value between the output value according to the training result of the second training sample and the output value of the already known training sample(S340). If the error value is under the critical value, the third sample is continuously trained by the first system. If the error value is above the critical value, the second system parameter according to the second sample is generated(S350). After detecting a state completing the training of the last sample(S360), a system self-organization procedure is ended by repeating the successive training of the sample for a constant time(S370).
展开▼