The adaptation of artificial neural network (ANN) weights using Genetic Algorithms (GAs) is a well known mechanism used to improve the performance of ANN to model manufacturing non-linear processes. GAs is capable on finding the best set of weights for the ANN and avoids being trapping in local minimum as in the case of Backpropagation (BP) learning algorithm. However, the learning process still suffer from being slow since the evolutionary processes based GAs takes tremendous amount of time especially if the modeling process is highly complex. To recover from this problem, Parallel Genetic Algorithms (PGAs) is proposed to evolve the ANN weight in the process of modeling the winding machine in the Aluminum factory. A comparison is made to show the effectiveness of the proposed approach between ANN-BP, ANN-GAs and ANN-PGAs. The obtained results are promising.

Login Required

Please login to access the full PDF of this article.

🔑 Author Login