EFFICIENT TIME-STEPPING FOR NUMERICAL INTEGRATION

doi:10.1137/21M1412682

EFFICIENT TIME-STEPPING FOR NUMERICAL INTEGRATION

2023-04-26

2

OA

AI

DOI：10.1137/21M1412682

Original

Original request for help

Save

Abstract

En 中文

Many problems in science and engineering require an efficient numerical approxima-tion of integrals or solutions to differential equations. For systems with rapidly changing dynamics, an equidistant discretization is often inadvisable as it results in prohibitively large errors or computa-tional effort. To this end, adaptive schemes, such as solvers based on Runge--Kutta pairs, have been developed which adapt the step size based on local error estimations at each step. While the classical schemes apply very generally and are highly efficient on regular systems, they can behave sub opti-mally when an inefficient step rejection mechanism is triggered by structurally complex systems such as chaotic systems. To overcome these issues, we propose a method to tailor numerical schemes to the problem class at hand. This is achieved by combining simple, classical quadrature rules or ODE solvers with data-driven time-stepping controllers. Compared with learning solution operators to ODEs directly, it generalizes better to unseen initial data as our approach employs classical numeri-cal schemes as base methods. At the same time it can make use of identified structures of a problem class and, therefore, outperforms state-of-the-art adaptive schemes. Several examples demonstrate superior efficiency. Source code is available at https://github.com/lueckem/quadrature-ML.

Keywords:

Key words

initial value problems

quadrature

time-stepping

machine learning

reinforcement learning

EFFICIENT TIME-STEPPING FOR NUMERICAL INTEGRATION

Abstract

Journal

Organization