A fast algorithm is presented for time-domain statistical analysis of large circuits with multiple stochastic parameters. The proposed technique is based on Numerical Inversion of Laplace Transform and does not require explicit presentation of the dynamic model in the form of set of differential equations. As a result, the moment vectors associated with frequency can be excluded when forming the moments subspace, leading to much smaller reduced models. In addition, the algorithm is highly parallelizable compared to time-stepping numerical integration techniques.

Additional Metadata
Persistent URL dx.doi.org/10.1109/SaPIW.2016.7496288
Conference 20th IEEE Workshop on Signal and Power Integrity, SPI 2016
Citation
Tao, Y. (Ye), Nouri, B. (Behzad), Nakhla, M.S, & Achar, R. (2016). Efficient time-domain variability analysis using parameterized model-order reduction. In 2016 IEEE 20th Workshop on Signal and Power Integrity, SPI 2016 - Proceedings. doi:10.1109/SaPIW.2016.7496288