Multi-parametric uncertainty quantification with a hybrid Monte-Carlo / polynomial chaos expansion FDTD method

The increasing complexity of microwave structures necessitates the quantification and analysis of fabrication process uncertainties that directly impact their performance, as a means of ensuring performance robustness. A particular technique that has emerged as a promising alternative to the time-consuming Monte-Carlo method for such analyses, is the polynomial chaos expansion (PCE). However, PCE-based methods suffer from a rapid increase in their computational cost with the number of random variables included in the analysis. This paper demonstrates that a hybrid Monte-Carlo/PCE technique can lead to the dramatic acceleration of its constituent methods, in the context of FDTD, effectively mitigating the “curse of dimensionality” and facilitating the multi-parametric uncertainty analysis of electromagnetic geometries.