by jro
6m715u
Volume 12 number 1
Pages: 1-10
Projeto e anlise de uma rede neural para resolver problemas de programao dinmica
Ivan N. da Silva, Lcia V. R. Arruda, Wagner C. do Amaral e Mrio E. Bordon
| Resumo: | Redes Neurais Artificiais so sistemas dinmicos que possuem altas taxas de computao por utilizarem um nmero elevado de elementos processadores simples com alta taxa de conectividade entre si. Redes neurais com conexes retroalimentadas fornecem um modelo computacional capaz de resolver vrios tipos de problemas de otimizao. Este artigo apresenta uma nova abordagem para resolver problemas de programao dinmica utilizando redes neurais artificiais. Mais especificamente, uma rede de Hopfield modificada desenvolvida cujos parmetros internos so computados utilizando a tcnica de subespao-vlido de solues. Estes parmetros garantem a convergncia da rede em direo aos pontos de equilbrio que representam as solues (no necessariamente timas) para o problema de programao dinmica. Resultados de simulaes so apresentados para validar a abordagem proposta. |
| Palavras Chave: | Redes neurais artificiais, programao dinmica, redes de Hopfield, otimizao de sistemas. |
| Abstract: | Analysis and Design of an Artificial Neural Networks for Solving Dynamic Programming Problems. Systems based on artificial neural networks have high computational rates due to the use of a massive number of simple processing elements and the high degree of connectivity between these elements. Neural networks with connections provide a computing model capable of solving a large class of optimization problems. This paper presents a novel approach for solving dynamic programming problems using artificial neural networks. More specifically, a modified Hopfield network is developed and its internal parameters are computed using the valid-subspace technique. These parameters guarantee the convergence of the network to the equilibrium points which represent solutions (not necessarily optimal) for the dynamic programming problem. Simulated examples are presented and compared with other neural networks. The results demonstrate that proposed method gives a significant improvement. |
| Keywords: | Artificial neural networks, dynamic programming, Hopfield networks, system optimization. |
PDF copy (168 kbytes)
| Click here to obtain
|
Last modifications: by jro |