Neural networks learning as a multiobjective optimal control problem.

Krawczak, Maciej. "Neural networks learning as a multiobjective optimal control problem.." Mathware and Soft Computing 4.3 (1997): 195-202. <http://eudml.org/doc/39108>.

@article{Krawczak1997,
abstract = {The supervised learning process of multilayer feedforward neural networks can be considered as a class of multi-objective, multi-stage optimal control problem. An iterative parametric minimax method is proposed in which the original optimization problem is embedded into a weighted minimax formulation. The resulting auxiliary parametric optimization problems at the lower level have simple structures that are readily tackled by efficient solution methods, such as the dynamic programming or the error backpropagation algorithm. The analytical expression of the partial derivatives of systems performance indices with respect to the weighting vector in the parametric minimax formulation is derived.},
author = {Krawczak, Maciej},
journal = {Mathware and Soft Computing},
keywords = {Redes neuronales; Teoría del aprendizaje; Control óptimo; Criterio minimax; Resolución de problemas; supervised learning; multilayer feedforward neural networks},
language = {eng},
number = {3},
pages = {195-202},
title = {Neural networks learning as a multiobjective optimal control problem.},
url = {http://eudml.org/doc/39108},
volume = {4},
year = {1997},
}

TY - JOUR
AU - Krawczak, Maciej
TI - Neural networks learning as a multiobjective optimal control problem.
JO - Mathware and Soft Computing
PY - 1997
VL - 4
IS - 3
SP - 195
EP - 202
AB - The supervised learning process of multilayer feedforward neural networks can be considered as a class of multi-objective, multi-stage optimal control problem. An iterative parametric minimax method is proposed in which the original optimization problem is embedded into a weighted minimax formulation. The resulting auxiliary parametric optimization problems at the lower level have simple structures that are readily tackled by efficient solution methods, such as the dynamic programming or the error backpropagation algorithm. The analytical expression of the partial derivatives of systems performance indices with respect to the weighting vector in the parametric minimax formulation is derived.
LA - eng
KW - Redes neuronales; Teoría del aprendizaje; Control óptimo; Criterio minimax; Resolución de problemas; supervised learning; multilayer feedforward neural networks
UR - http://eudml.org/doc/39108
ER -