Tuning of the structure and parameters of neural network using an improved genetic algorithm

Abstract

Presents the tuning of the structure and parameters of a neural network using an improved genetic algorithm (GA). The improved GA is implemented by floating-point arithmetic. The processing time of the improved GA is faster than that of the GA implemented by binary number as coding and decoding are not necessary. By introducing new genetic operators to the improved GA, it is also shown that the improved GA performs better than the traditional GA based on some benchmark test functions. A neural network with switches introduced to links is proposed. By doing this, the proposed neural network can learn both the input-output relationships of an application and the network structure. Using the improved GA, the structure and the parameters of the neural network can be tuned. An application example on sunspot forecasting is given to show the merits of the improved GA and the proposed neural network.