Rooted tree analysis of Runge-Kutta methods with exact treatment of linear terms

Authors:
S. Koikari
Affiliations:
Department of Mechanical Engineering, Graduate School of Engineering, Kyoto University, Yoshida-Honmachi, Sakyo-ku, Kyoto 606-8501, Japan
Venue:
Journal of Computational and Applied Mathematics
Year:
2005

Citing 13
Cited 1

Pseudo-Runge-Kutta Processes

Publications of the Research Institute for Mathematical Sciences
A method for exponential propagation of large systems of stiff nonlinear differential equations

Journal of Scientific Computing
Krylov methods for the incompressible Navier-Stokes equations

Journal of Computational Physics
A new class of time discretization schemes for the solution of nonlinear PDEs

Journal of Computational Physics
Pseudo-Runge-Kutta Methods Involving Two Points

Journal of the ACM (JACM)
Spectral methods in MatLab

Spectral methods in MatLab
A quasi-steady state solver for the stiff ordinary differential equations of reaction kinetics

Journal of Computational Physics
An exponential method of numerical integration of ordinary differential equations

Communications of the ACM
Exponential time differencing for stiff systems

Journal of Computational Physics
Additive Runge-Kutta schemes for convection-diffusion-reaction equations

Applied Numerical Mathematics
Adaptive solution of partial differential equations in multiwavelet bases

Journal of Computational Physics
Fourth-Order Time-Stepping for Stiff PDEs

SIAM Journal on Scientific Computing
Generalized integrating factor methods for stiff PDEs

Journal of Computational Physics

Beyond conventional Runge-Kutta methods in numerical integration of ODEs and DAEs by use of structures and local models

Journal of Computational and Applied Mathematics

Quantified Score

Hi-index	7.30

Visualization

Abstract

We investigate a class of time discretization schemes called "ETD Runge-Kutta methods," where the linear terms of an ordinary differential equation are treated exactly, while the other terms are numerically integrated by a one-step method. These schemes, proposed by previous authors, can be regarded as modified Runge-Kutta methods whose coefficients are matrices instead of scalars. From this viewpoint, we reexamine the notion of consistency, convergence and order to provide a mathematical foundation for new methods. Applying the rooted tree analysis, expansion theorems of both the strict and numerical solutions are proved, and two types of order conditions are defined. Several classes of formulas with up to four stages that satisfy the conditions are constructed, and it is shown that the power series of matrices, employed as their coefficients, can be determined using the order conditions.