Tuning of the Structure and Parameters of a Neural Network Using a Good Points Set Evolutionary Strategy

In this paper, a good points set-evolutionary strategy (GPSES) is applied to solve the problem of tuning both network structure and parameters of a feedforward neural network. Good point set (GPS) is a concept in number theory. To overcome the deficiency of orthogonal design to handle optimization problems, this paper presents a method that incorporate GPS principle to enhance the crossover operator of the evolution strategy can make the resulting evolutionary algorithm more robust and statically sound. The presented GPSES approach is effectively applied to solve the examples on forecasting the sunspot numbers. The numbers of hidden nodes and the links of the feedforward neural network are chosen by increasing them from small numbers until the learning performance is good enough. As a result, a partially connected feedforward neural network can be obtained after tuning. This implies that the cost of implementation of the neural network can be reduced. Experiment results show the efficiency of our methods.