Updated Kriging-Assisted Shape Optimization of a Gravity Dam

The numerical simulation of the optimal design gravity dams is computationally expensive. Therefore, a new optimization procedure presented in this study to reduce computational cost for determining shape dam. Optimization was performed using combination genetic algorithm (GA) and an updated Kriging surrogate model (UKSM). First, (KSM) constructed with small sample set. Second, minimizing predictor strategy used add samples region interest update KSM each updating cycle until process converged. Third, existing dam demonstrate effectiveness GA–UKSM. solution obtained GA–UKSM compared that GA–KSM. results revealed required only 7.53% total number simulations by GA–KSM achieve similar results. Thus, can significantly improve efficiency. method adopted be as reference dams.