A simplified model for vacancy type and self-interstitial-type defects was considered. The problem of the model was explained and a neural network reinforced improvement was adapted to the model. The improved analytical model was compared with the finite volume technique based numerical solution. It was observed that the model gained better accuracy and validity with the aid of a neural network.

Neural Network Reinforced Point Defect Concentration Estimation Model for Czochralski-Grown Silicon Crystals. M.Avci, S.Yamacli: Mathematical and Computer Modelling, 2010, 51[7-8], 857-62