Wednesday, November 27th, 2019. Time: 15h.
Place: O.C. Zienkiewicz Conference Room, C1 Building, UPC Campus Nord, Barcelona
Dealing with uncertainties is necessary in many different science and engineering scenarios, and this demanding necessity can be properly assessed nowadays, thanks to the development of computational capabilities.
In this scenario, current solvers must consider both statistical convergence and computational efficiency when developing strategies.
Monte Carlo is the reference method when dealing with stochastic problems. Unfortunately, the convergence of this well-known algorithm is slow, making it prohibitively expensive for complex engineering problems.
The aim of this talk is to present a new way of increasing Monte Carlo performance: by developing an asynchronous framework for this class of algorithms.
The asynchronous Monte Carlo framework increases the computational efficiency, while keeping the same statistical reliability and accuracy of state-of-art algorithms.
Comparisons with both determistic and hierarchy-adaptive Monte Carlo algorithms have been performed, to verify the accuracy and the improvements of the proposed framework.
 Dadvand, P., Rossi, R., & Oñate, E. (2010). An object-oriented environment for developing finite element codes for multi-disciplinary applications. Arch Comput Methods Eng, 17(3), 253–297.
 Pisaroni, M., Nobile, F., & Leyland, P. (2017). A Continuation Multi Level Monte Carlo (C-MLMC) method for uncertainty quantification in compressible inviscid aerodynamics. Comput Method Appl M, 326, 20–50.
Riccardo Tosi is a Mathematical Engineering graduated (dottore magistrale) from the University of Padova, Italy (2018). He is a PhD student at CIMNE in the KratosMultiphysics research group from September 2018. His research work is centred on Uncertainty Quantification, with applications to wind engineering problems.