Numerical simulations of stochastic inverse problems

题    目:  Numerical simulations of  stochastic inverse  problems

报告人:  Wenju Zhao

Department of Scientific Computing

Florida State University, USA

地    点: 山东大学中心校区知新楼B1248

时    间: 2017年1月9日上午10:30

 

This research intends to present a numerical method for solving a real application problem in the fluid dynamic fields characterized by a stochastic inverse models. The sparse observations, noisy measurements, highly nonlinear, and non-Gaussian properties of real problems lead to complicated representations of the stochastic inverse solutions. To incorporate the observation data and PDE model, a probabilistic solution is constructed by the objective Bayesian inferences framework. To efficiently deal with high-dimensional properties and accelerate the numerical simulations, some advanced computational techniques in the fields of uncertainty quantity, PDE constrained optimal control, etc. are combined together. Finally, the numerical tests are presented.