作者: G.G. Yen , Haiming Lu
关键词: Recurrent neural network 、 Feedforward neural network 、 Artificial intelligence 、 Time delay neural network 、 Nervous system network models 、 Deep learning 、 Neural gas 、 Probabilistic neural network 、 Physical neural network 、 Computer science
摘要: In this paper, we propose a novel genetic algorithm based design procedure for multi-layer feedforward neural network. Hierarchical is used to evolve both network topology and parameters. Compared with traditional designs network, the proposed hierarchical approach addressed several deficiencies highlighted in literature. A multi-objective function herein optimize performance of evolved Two benchmark problems are successfully verified proves be competitive or even superior back-propagation Mackey-Glass chaotic time series prediction.