Tuning the structure and parameters of a neural network using cooperative binary-real particle swarm optimization

In this paper, a cooperative binary-real particle swarm optimization is applied to tune the structure and parameters of a neural network. A neural network with switches of its links, which is used to decide whether there is a link between two neurons or not, is introduced firstly. Thus, the structure of a neural network can be decided by the switches. A cooperative binary-real particle swarm optimization algorithm is utilized to find the compact structures and optimal parameters of the proposed neural network. The number of hidden nodes of the neural network is increased from a small number until its learning ability is achieved. The simulation experiments indicate that the proposed approach can obtain better results than the existing approaches in recent literature.