The article discusses how Kriging interpolation is used in simulations to predict output values based on observed inputs. Kriging provides accurate predictions in deterministic simulations and is becoming more popular in discrete-event simulations. Compared to traditional regression analysis, Kriging often gives better global predictions by giving more weight to nearby observations. To create the Kriging model, space-filling designs like latin hypercube sampling are used. The paper also introduces new customized designs for better accuracy.