Rank-adaptive tensor methods for high-dimensional nonlinear PDEs

12/10/2020
by   Alec Dektor, et al.
0

We present a new rank-adaptive tensor method to compute the numerical solution of high-dimensional nonlinear PDEs. The new method combines functional tensor train (FTT) series expansions, operator splitting time integration, and a new rank-adaptive algorithm based on a thresholding criterion that limits the component of the PDE velocity vector normal to the FTT tensor manifold. This yields a scheme that can add or remove tensor modes adaptively from the PDE solution as time integration proceeds. The new algorithm is designed to improve computational efficiency, accuracy and robustness in numerical integration of high-dimensional problems. In particular, it overcomes well-known computational challenges associated with dynamic tensor integration, including low-rank modeling errors and the need to invert the covariance matrix of the tensor cores at each time step. Numerical applications are presented and discussed for linear and nonlinear advection problems in two dimensions, and for a four-dimensional Fokker-Planck equation.

READ FULL TEXT

Authors

page 15

page 17

page 19

07/19/2020

Dynamic tensor approximation of high-dimensional nonlinear PDEs

We present a new method based on functional tensor decomposition and dyn...
08/01/2020

Step-truncation integrators for evolution equations on low-rank tensor manifolds

We develop a new class of algorithms, which we call step-truncation meth...
07/08/2019

Deep splitting method for parabolic PDEs

In this paper we introduce a numerical method for parabolic PDEs that co...
09/07/2021

A low-rank tensor method to reconstruct sparse initial states for PDEs with Isogeometric Analysis

When working with PDEs the reconstruction of a previous state often prov...
10/02/2014

Scalable Nonlinear Learning with Adaptive Polynomial Expansions

Can we effectively learn a nonlinear representation in time comparable t...
07/12/2019

Dynamically orthogonal tensor methods for high-dimensional nonlinear PDEs

We develop new dynamically orthogonal tensor methods to approximate mult...
02/20/2022

Fast high-dimensional integration using tensor networks

The design and application of regression-free tensor network representat...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.