Study On Kriging Coupled Model Method Based On Radial Basis Neural Network

In recent years,with the rapid improvement of computing power,numerical simulation plays an increasingly important role in scientific research and engineering design.However,due to the inevitable approximation,simplification and human factors in numerical simulation,the reliability of numerical simulation results has become an issue of increasing concern.Uncertainty quantification(UQ)is an emerging research direction of computational mathematics in recent years.Its function is to quantitatively characterize the degree of complex processes reflected by simulation results.It is difficult to carry out large sample calculation due to the large amount of computation in numerical simulation for many problems.So people put forward the surrogate model method.In order to further improve the accuracy of the surrogate model,based on RBF neural network and the advantages and disadvantages of Kriging surrogate model analysis,this paper proposes a Kriging coupled model based on RBF neural network,the method of numerical experiments conducted by three analytic function,and verify the numerical results show that our proposed coupled method is effective.