Modified Sequential Kriging Optimization for Multidisciplinary Complex Product Simulation

AbstractDirecting to the high cost of computer simulation optimization problem, Kriging surrogate model is widely used to decrease the computation time. Since the sequential Kriging optimization is time consuming, this article extends the expected improvement and put forwards a modified sequential Kriging optimization (MSKO). This method changes the twice optimization problem into once by adding more than one point at the same time. Before re-fitting the Kriging model, the new sample points are verified to ensure that they do not overlap the previous one and the distance between two sample points is not too small. This article presents the double stopping criterion to keep the root mean square error (RMSE) of the final surrogate model at an acceptable level. The example shows that MSKO can approach the global optimization quickly and accurately. MSKO can ensure global optimization no matter where the initial point is. Application of active suspension indicates that the proposed method is effective