Combine Gaussian Bare-Bones Differential Evolution with Hybrid Gaussian Mutation Strategy in Optimization of GM(1,1)

In order to improve the forecasting accuracy of grey model, this study attempts to apply Gaussian bare-bones differential evolution to optimize the background value of grey model. In which a hybrid-type Gaussian mutation, hybridizing two Gaussian mutation strategies into one expression, is proposed to enhance the global searching ability of GBDE. The proposed hybrid-type mutation strategy is also almost parameter free. Experimental results show that the proposed Gaussian bare-bones differential evolution -based GM(1,1) has better fitting and forecasting accuracy than the original GM(1,1) and the genetic algorithm based GM(1,1).