Efficient Time-Stepping for Numerical Integration Using Reinforcement Learning
From MaRDI portal
Initial value problems, existence, uniqueness, continuous dependence and continuation of solutions to ordinary differential equations (34A12) Numerical quadrature and cubature formulas (65D32) Numerical methods for initial value problems involving ordinary differential equations (65L05) Multistep, Runge-Kutta and extrapolation methods for ordinary differential equations (65L06) Learning and adaptive systems in artificial intelligence (68T05)
Abstract: Many problems in science and engineering require an efficient numerical approximation of integrals or solutions to differential equations. For systems with rapidly changing dynamics, an equidistant discretization is often inadvisable as it either results in prohibitively large errors or computational 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-optimal 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 generalises better to unseen initial data as our approach employs classical numerical 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.
Recommendations
- Personalized algorithm generation: a case study in learning ODE integrators
- A machine learning framework for data driven acceleration of computations of differential equations
- Time-step selection algorithms: adaptivity, control, and signal processing
- Automatic control and adaptive time-stepping
- scientific article; zbMATH DE number 431523
Cites work
- \({\mathcal Q}\)-learning
- A family of embedded Runge-Kutta formulae
- A Fourier integrator for the cubic nonlinear Schrödinger equation with rough initial data
- Data-driven approximation of the Koopman generator: model reduction, system identification, and control
- DGM: a deep learning algorithm for solving partial differential equations
- Geometric Numerical Integration
- scientific article; zbMATH DE number 1796810 (Why is no real title available?)
- scientific article; zbMATH DE number 940566 (Why is no real title available?)
- scientific article; zbMATH DE number 3802727 (Why is no real title available?)
- scientific article; zbMATH DE number 1865611 (Why is no real title available?)
- Machine learning for fast and reliable solution of time-dependent differential equations
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Reinforcement learning. An introduction
- Solving Ordinary Differential Equations I
- The MATLAB ODE Suite
- Vectorized adaptive quadrature in MATLAB
Cited in
(4)- NySALT: Nyström-type inference-based schemes adaptive to large time-stepping
- Personalized algorithm generation: a case study in learning ODE integrators
- Learning Robust Marking Policies for Adaptive Mesh Refinement
- Reinforcement learning for adaptive time-stepping in the chaotic gravitational three-body problem
This page was built for publication: Efficient Time-Stepping for Numerical Integration Using Reinforcement Learning
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6039265)
