Paper Title
Optimization of ANN Weights for an Industrial Process Modeling Using Parallel GAs

Abstract
Winding systems are complex industrial processes which have a vast range of critical applications such as wide-plant processes that contain web conveyance and rolling steel industries. Modeling winding systems is a challenging task since it is multivariable, nonlinear and time variant. Genetic Algorithms (GAs) was effectively applied to optimize the weights of Artificial Neural Networks (ANNs) such that the learning process did not trap at a local minimum by the use of backpropagation learning algorithm. Here, Parallel Genetic Algorithm (PGA) was used to optimize ANN weights to model the dynamics of a winding machine. The proposed method increased the speed of the learning process during weight optimization. A comparison between GAs and PGA for ANN weight optimization is shown. The results are promising in the case of PGA. Keywords - Artificial neural networks; parallel genetic algorithm; style; winding system