Comparison of the radial basis function and ordinary kriging model in the design of dynamic systems

Metamodels such as radial basis functions (RBF) and Kriging models are becoming popular techniques for approximating complex mechanistic models. In the case of robust design where the design space is investigated for the best design parameters of the design variables, metamodels offer a fast and simple method since they can be easily applied to Monte Carlo methods. Most research in the area of metamodels compares the performance of the metamodel in approximating the existing mechanistic model. Much research looks at the performance of regression models with either Kriging or RBFs in their ability to approximate the mechanistic model. This paper will observe the performance of the radial basis function (RBF) with Kriging models in terms of the sustainable design of a dynamic system where design is performed using a Monte Carlo simulation (MCS).