Stocastic photonics: Tools and approaches for the analysis and optimization of integrated circuits

We demonstrate the application of several stochastic tools to the design of photonic integrated circuit. Monte Carlo analysis based on real measured statistics of the parameters is used to investigate of the impact of fabrication tolerances on the performance of a switch, including average power required for trimming. Polynomial-chaos-based models are used to optimize a filter design in order to minimize fluctuations of the bandwidth due to fabrication uncertainty. Sensitivity analysis performed through elementary effect test and variance based approach is used to identify the most critical parameters determining bandwidth fluctuations.