Xueyu Zhu - Colloquium Speaker
We present an overview of the recent developments of multi-fidelity algorithms for uncertainty quantification problems. The method combines the computational efficiency of low-fidelity models with high accuracy of high-fidelity models. We shall illustrate the advantages of the method via a set of more comprehensive benchmark examples. Both the mathematical justification and numerical implementation are discussed. The method is able to produce accurate estimates of the solution moments using a very small number of expensive high-fidelity simulations. Various numerical tests in different applications fields are provided to demonstrate the performance of the method.