applied math seminar
Gianluca Geraci, Sandia National Laboratories
Monday, April 8, 2019 - 3:30pm
TITLE: Multilevel/Multifidelity sampling strategies for Uncertainty Quantification: a generalized interpretation through the lens of the approximate control variate framework Speaker: Gianluca Geraci Co-authors: Alex A. Gorodetsky, Michael S. Eldred and John D. Jakeman Abstract In the last decades, the advancements in both computer hardware/architectures and scientific computing algorithms enabled engineers and scientists to more rapidly study and design complex systems by heavily relaying on numerical simulations. Uncertainty Quantification (UQ) evolved as a task within the most comprehensive Verification and Validation framework which aims at obtaining truly predictive numerical simulations. Despite the recent efforts and successes in advancing the algorithms’ efficiency, the combination of a large set of uncertainty parameters (often correlated to the complexity of the numerical/physical assumptions) and the lack of regularity of the system's response still represents a formidable challenge for UQ. One of the possible ways of circumventing these difficulties is to rely on sampling-based approaches. Sampling based approaches are well known for their robustness and simplicity. However, these methods are also characterized by a slow rate of convergence which makes them often unaffordable for realistic problems. In this talk we summarize our recent efforts in investigating novel ways of increasing the efficiency of these methods by resorting to multilevel and multifidelity concepts which, under certain assumptions, can be generalized in the so-called approximate control variate framework. Additionally, increasing the efficiency of these techniques by identifying common structures amongst models will also be discussed. Several numerical examples will be presented ranging from simple verification cases up to more realistic engineering problems. Bio Gianluca Geraci is a Senior Member of Technical Staff in the Optimization and Uncertainty Quantification department of the Sandia National Laboratories in Albuquerque, New Mexico. In 2010 he received a Master degree in Aeronautical Engineering from Politecnico di Milano, Italy with a major in aerodynamics and a thesis on hybrid finite element/finite volume schemes for compressible flows in curvilinear coordinates. In 2013 he received a PhD in Applied Mathematics and Scientific Computing with a thesis on intrusive multiresolution UQ schemes carried out at the French Institute for Research in Computer Science and Automation (INRIA) under the direction of Prof. Remi Abgrall. After his PhD he served as PostDoctoral Fellow at the Center for Turbulence Research of the Stanford University under the direction of Prof. Gianluca Iaccarino working on UQ analysis of particle-laden turbulent flows in radiative environments. His main interests are compressible fluid dynamics, uncertainty quantification and optimization under uncertainty.
Contact Name: Pavel Lushnikov