Título: Algoritmo Rápido de Treinamento de Redes Neurais para Sistemas de Controle em Tempo Real
Autores: Silva, L. E. Borges da; Silva, A. P. Alves da; Lambert-Torres, G.; Ribeiro, E. R.
Resumo: The long time generally required for training ANN has been a critical problem for the utilization of this technology in real-time. The generalized delta rule with backward error propagation (backpropagation) has been the most used training algorithm for ANN in control applications. However, despite many attempts to improve the performance of this learning algorithm, there is still no efficient and reliable method to train a multilayer perceptron. The main problem is high nonlinearity of the error function. This paper presents an alternative method for load information into a multilayer perseptron. The idea is to use successive quadratic approximations for the error function, until getting into the proximity of the global minimum. This technique was applied to Control System in order to obtain an adaptive behaviour.
Código DOI: 10.21528/CBRN1994-040
Artigo em PDF: CBRN1994-paper40.pdf
Arquivo BibTex: CBRN1994-paper40.bib