Incremental Kriging Model Rebuilding Method and its Application in Efficient Global Optimization

In efficient global optimization(EGO) algorithm,the time of rebuilding the Kriging model increases rapidly with the increasing of samples' size,and premature convergence may exist when the range of the objective function is too large.To conquer these problems,an incremental Kriging method(IKM) and the improved EGO algorithm are proposed.The inversion of the correlation matrix and the new data points are manipulated to get the coefficients of the Kriging model in IKM,while coefficients of correlation function are optimized and the inversion of new correlation matrix is directly calculated.Stopping criteria on expected improvement,response value and argument are used in the improved EGO algorithm.The experimental results demonstrate that IKM greatly reduces the time of modelling with little loss of accuracy and the improved EGO method has higher efficiency and better stability.