Artificial Neural Network (ANN) models are applied to both simulated and experimental data from the ASDEX
Upgrade tokamak. The results are compared in detail to those obtained from the real time version of the Function Parameterization (FP) technique currently in use for the feedback control of the plasma discharge and for after-shot analysis. Though the on-line evaluation algorithm for both methods is quite similar, the choice of nonlinear models and the strategies for determining the representation coefficients are very different. Both methods can be advantageously used to solve problems of plasma identification.