Global optimization of electromagnetic devices combining Latin hypercube sampling experiment and adaptive response surface method

This paper presents a global optimization algorithm of electromagnetic devices using an adaptive response surface approximation of the objective function with sampling strategy based on design of experiments. In the adaptive approximation, an optimal Latin hypercube design based on multi-objective Pareto optimization is developed to obtain the sampling data in the design variable space. The proposed method is verified with an analytic test function and is applied to optimal design of electromagnetic devices.