An Improved Grey Neural Network Model with GA Optimization

Although the grey forecasting model has been successfully employed in various fields and demonstrated promising results, its performance still could be improved. In order to improve the fitting capability of grey model, an improved grey neural network model with GA optimization is proposed in this paper. To avoid the premature convergence and inbreeding, an improved GA (IGA) is proposed in this paper. This IGA is used to improve the grey neural network. Binary encoding and random uniform distribution are employed in the course of initializing population with the purpose of increasing the population diversity, which speed up optimizing. Adaptive probability of crossover and mutation are used to avoid the population getting into local optimum and they change with the change of fitness of population. The initial condition of grey model is not consistent with theory. The improved GA is used to optimize the paramerter of the improved GM proposed in this paper. Simulation experiment indicates the improved model is effective in increasing prediction precision.