Conference article

Kriging model is a commonly interpolate approximation method which is widely used in the computer simulation in the past decade. The fitting accuracy is one of the fundamental problems in the research of kriging model, which can be summarized in two aspects, the accurate estimation of model’s parameters and approximate form selection of kriging model. In order to solve the existed problems, an improved parameter estimation method of kriging model base on differential evolution (DE) algorithm is set out in the present paper. Firstly, establish the objective function of DE algorithm depends on the estimation of the model’s accuracy, and get the optimum solution of model’s parameters under the initial condition. Then, a variety of regression function and correlation function in kriging models are selected to compare the fitting accuracy. Finally, the simulation case for outer ballistic data on electromagnetic railgun is examined to determine whether the improved method has priority over traditional one in the approximation accuracy.

Kriging model, DE algorithm, approximation accuracy, EM railgun