A Global Robust Optimization Using Kriging Based Approximation Model

The current trend of design methodologies is to make engineers objectify or automate the decision-making process. Numerical optimization is an example of such technologies but it may produce uncontrollable uncertainties. To increase manageability of such uncertainties, the Taguchi method, reliability-based optimization and robust optimization are commonly being used. The main functional requirement of a mechanical system is to obtain the target performance with maximum robustness. In this research, a design procedure for global robust optimization is developed using kriging and global optimization approaches. Robustness is determined by kriging model to reduce a number of real functional calculations. The simulated annealing algorithm of global optimization methods is adopted to determine the global robust optimum of a surrogate model. As the postprocess, the global optimum is further refined by applying the first-order second-moment approximation method. Mathematical problems and the MEMS design problem are investigated to show the validity of the proposed method.