A class of accelerators for BDF methods
backward differentiation formula methodBDF methodsconvergence accelerationerror estimateimplicit Euler methodnumerical examplesorder accelerationstability
Nonlinear ordinary differential equations and systems (34A34) Numerical methods for initial value problems involving ordinary differential equations (65L05) Multistep, Runge-Kutta and extrapolation methods for ordinary differential equations (65L06) Stability and convergence of numerical methods for ordinary differential equations (65L20) Error bounds for numerical methods for ordinary differential equations (65L70)
It is shown how the order of the \(k\)-step backward differentiation formula (BDF) method of order \(k\) can be accelerated by one by adding a correction term containing the value of the Jacobian at the update point. This does not require extra function evaluations or LU-factorizations and preserves the stability properties of the original method. An example accelerating the order of the implicit Euler method to two and provision of a local error estimate is given. Some brief numerical tests suggest that this approach is a promising one.
- scientific article; zbMATH DE number 4045044
- scientific article; zbMATH DE number 221341
- scientific article; zbMATH DE number 5012790
- scientific article; zbMATH DE number 4161339
- On the acceleration of the Barzilai-Borwein method
- Parallel implementation of BVM methods
- Acceleration schemes for the method of alternating projections
- scientific article; zbMATH DE number 3915507
- Accelerated bundle level methods with inexact oracle
This page was built for publication: A class of accelerators for BDF methods
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1294071)
