Numerical approximation of high-dimensional Fokker-Planck equations (C08)

点击翻译按钮获取中文摘要

The goal of this project is the development of numerical methods for the error-controlled approximation of Fokker-Planck equations in high dimensions, with particular focus on equations describing dilute polymer solutions and solutions with active particles. Two different classes of approximations are considered: sparse product basis expansions parameterized by low-rank tensor decompositions, and approximations by sums of high-dimensional Gaussian functions.