Ahmad Jafarian - Department of Mathematics, Urmia Branch, Islamic Azad University, Urmia, Iran
Safa Measoomy Nia - Department of Mathematics, Urmia Branch, Islamic Azad University, Urmia, Iran
Raheleh Jafari - Department of Mathematics, science and research Branch, Islamic Azad University, Arak, Iran

Artificial neural networks have the advantages such as learning,adaptation, fault-tolerance, parallelism and generalization. This paper mainlyintends to offer a novel method for finding a solution of a fuzzy equation thatsupposedly has a real solution. For this scope, we applied an architecture offuzzy neural networks such that the corresponding connection weights are realnumbers. The suggested neural net can adjust the weights using a learningalgorithm that based on the gradient descent method. The proposed method isillustrated by several examples with computer simulations.

Fuzzy equations, Fuzzy feed-forward neural network (FFNN), Cost function,Learning algorithm