915 Recent Advances in Numerical Analysis and Algorithms for Uncertainty Quantification and Its Applications
Guannan Zhang, Oak Ridge National Laboratory
Clayton Webster, University of Tennessee
Mathematical analysis and computational simulations for engineered complex systems are often affected by many sources of uncertainty, such as limited data availability or accuracy, approximations in modeling, and computational processing. Uncertainty quantification is the science of quantitative characterization and reduction of uncertainties in both computational and real world applications. This mini-symposium focuses on the fundamental problem of how to accurately approximate solutions of both forward and inverse complex systems with random input data. Predicting the behavior of complex phenomena relies on constructing solutions in terms of high dimensional spaces, particularly in the case when the input data (coefficients, forcing terms, initial and boundary conditions, geometry) are affected by large amounts of uncertainty. This mini-symposium aims at exploring breakthroughs in sparse polynomial approximation, multilevel methods, compressed sensing, model calibration and data-driven reduced order modeling, Bayesian inference and stochastic control/optimization.