A neural network inversion approach to electromagnetic device design

In this paper we present a new model that employs, in a natural and effecti ve way, a neural network inversion algorithm providing a solution of the el ectromagnetic device design problem, The model combines two types of artifi cial neural networks and applies a neural network inversion algorithm, Furt her development of this model can propose an efficient solution to the elec tromagnetic device design problem. The model is applied to the magnetic sti mulation coil design as wed as gradient coil design. The results obtained s how the effectiveness of the proposed model.