Generalized anisotropic polynomial chaos approach for expedited statistical analysis of nonlinear radio-frequency (RF) circuits

In this paper, a generalized anisotropic polynomial chaos approach for the fast statistical analysis of nonlinear radio-frequency (RF) circuits is proposed. This approach relies on a high dimensional model representation (HDMR) formulation in order to determine the maximum degree of the PC expansion along each random dimension separately. The resultant anisotropic multidimensional PC expansion of the circuit response is significantly sparse compared to the conventional isotropic expansion, thereby requiring substantially smaller time and memory cost for construction without significant loss of accuracy. The validity of the proposed methodology is demonstrated using a RF low noise amplifier example.