Multidimensional Uncertainty Quantification of Microwave/RF Networks Using Linear Regression and Optimal Design of Experiments

In this paper, a nonintrusive polynomial chaos (PC) approach for the fast multidimensional uncertainty quantification of microwave/RF networks is presented. The key feature of this paper is the development of a linear regression methodology based on sparse D-optimal design of experiments to accurately evaluate the PC coefficients of the network responses. A greedy search algorithm to identify the sparse set of D-optimal design of experiments from within a multidimensional random space has been presented. Additional numerical approaches to further accelerate the search algorithm for high-dimensional random spaces have also been developed.