An Automated Framework for Variability Analysis for Integrated Circuits Using Metaheuristics

This work aims to analyze the variability of integrated circuits and systems. An automated framework is presented for variability analysis that exploits the metaheuristic optimization techniques. The efficacy of the proposed approach is demonstrated by two case studies—one is the estimation of variability in phase noise in RF CMOS LC tank oscillators (frequency domain analysis) and the other is the estimation of variability in the differential output signal of a current mode driver (time-domain analysis). The proposed approach is investigated and validated by comparing the results from the traditional Monte Carlo simulations and the ordinary least-squares-based polynomial chaos expansion. A significant gain in the computational time is reported while maintaining accuracy in the results. The proposed methodology is not just limited to variability analysis applications but also can be used to solve the circuit optimization problems.