Multilevel Monte Carlo Methods with Smoothing

Learn about multilevel Monte Carlo (MLMC) methods enhanced with smoothing techniques for solving partial differential equations with uncertain parameters in this 53-minute mathematical lecture. Explore how to quantify uncertainty in model outputs when input parameters cannot be determined accurately by modeling them as stochastic processes. Discover the application of circulant embedding methods for coefficient sampling and understand how integrating smoothing techniques into the circulant embedding method significantly improves computational complexity. Master the approach that enables independent selection of the coarsest mesh on the first MLMC level regardless of the correlation length of the random field's covariance function, resulting in substantial computational cost savings for uncertainty quantification problems.