Research on an Accuracy Optimization Algorithm of Kriging Model Based on a Multipoint Filling Criterion

The optimization method based on the surrogate model has been widely used in simulation and calculation process of complex engineering models. However, this process, low accuracy computational efficiency always an urgent problem that needs to be solved. Aimed at problem, combined with two characteristics global search local detection, a filling criterion multiple points is firstly proposed named maximum expected improvement & minimizing predicted objective function root mean squared error (EI&MP&RMSE) paper. Furthermore, procedure EI&MP&RMSE concluded. Meanwhile, classical one-dimensional two-dimensional functions are applied verify method. difference square under different infill criteria analyzed. As expected, it shows can effectively improve reduce number iterations. It extensive practicability serviceability for structures.