Applying Kriging model for global design optimization
Xiang-jun Yu · Chinese journal of construction machinery · 2006
Based on the original algorithm, the Kriging model is employed as an approximation model for Bayesian analysis. This algorithm is adapted to resolve the complicated yet timeconsuming global design optimization problems.In this study, the initial samples are selected using the experimental design method, whereas the Kringing model is updated based on new sample data via the infill sample criterion. Thereafter, a global optimum can be obtained upon completion of sufficient iterations. It is proven that, compared with traditional optimization algorithms, this approach posesses high computational speed and precision.