State of the Art and Recent Progress in Uncertainty Quantification for Electronic Systems (i.e. Variation-Aware or Stochastic Simulation)
Friday, July 31, 2015 - 8:30am - 9:10am
ARMS Room 1010
On-chip and off chip fabrication process variations have become a major concern in today’s electronic systems design since they can significantly degrade systems’ performance. In order to predict and quantify such degradation, existing commercial circuit and MEMS simulators mostly rely only on the well known Monte Carlo algorithm. However during the last decade a large variety of more sophisticated and efficient alternative approaches have been proposed to accelerate such critical task. This talk will first review the state of the art of most modern intrusive and sampling-based uncertainty quantification techniques. It will then conclude showing in particular how parameterized model reduction, low-rank polynomial chaos expansions and hierarchical techniques can be used to accelerate most uncertainty quantification tools and to handle the curse of dimensionality.