ANN （artificial neural network） is a technique successfully employed in many applications on several research fields. An appropriate configuration for neural networks is a tedious task, and it often requires the kno...ANN （artificial neural network） is a technique successfully employed in many applications on several research fields. An appropriate configuration for neural networks is a tedious task, and it often requires the knowledge of an expert on the application. In this paper, a technique for automatic configuration for two types of neural networks is presented. The multilayer perceptron and recurrent Elman are the neural networks used here. The determination of optimal parameters for the neural network is formulated as an optimization problem, solved with the use of meta-heuristic MPCA （multiple particle collision algorithm）. The self-configuring networks are applied to perform data assimilation.展开更多
文摘ANN （artificial neural network） is a technique successfully employed in many applications on several research fields. An appropriate configuration for neural networks is a tedious task, and it often requires the knowledge of an expert on the application. In this paper, a technique for automatic configuration for two types of neural networks is presented. The multilayer perceptron and recurrent Elman are the neural networks used here. The determination of optimal parameters for the neural network is formulated as an optimization problem, solved with the use of meta-heuristic MPCA （multiple particle collision algorithm）. The self-configuring networks are applied to perform data assimilation.
