Hybrid Artificial Bee Colony algorithm for neural network training

• Published in: 2011 IEEE Congress of Evolutionary Computation (CEC) pp. 84 - 88
• Main Authors: Ozturk, Celal, Karaboga, Dervis
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2011
• Subjects: Approximation algorithms; Artificial bee colony algorithm; Artificial neural networks; Evolutionary computation; Hybrid algorithms; Levenberq-Marquardt algorithm; Neural network training; Neurons; Simulated annealing; Training
• ISBN: 1424478340; 9781424478347
• ISSN: 1089-778X
• DOI: 10.1109/CEC.2011.5949602

A hybrid algorithm combining Artificial Bee Colony (ABC) algorithm with Levenberq-Marquardt (LM) algorithm is introduced to train artificial neural networks (ANN). Training an ANN is an optimization task where the goal is to find optimal weight set of the network in training process. Traditional training algorithms might get stuck in local minima and the global search techniques might catch global minima very slow. Therefore, hybrid models combining global search algorithms and conventional techniques are employed to train neural networks. In this work, ABC algorithm is hybridized with the LM algorithm to apply training neural networks.