The Effective Stability of Adaptive Timestepping ODE Solvers

We consider the behavior of certain adaptive timestepping methods, based upon embedded explicit Runge--Kutta pairs, when applied to dissipative ODEs. It has been observed numerically that the standard local error controls can impart desirable stability properties, but this has been rigorously verified only for very special, low-order, Runge--Kutta pairs. The rooted-tree expansion of a certain quadratic form, central to the stability theory of Runge--Kutta methods, is derived. This, together with key assumptions on the sequence of accepted time-steps and the local error estimate, provides a general explanation for the observed stability of such algorithms on dissipative problems. Under these assumptions, which are expected to hold for "typical" numerical trajectories, two different results are proved. First, for a large class of embedded Runge--Kutta pairs of order $(1,2)$, controlled on an error-per-unit-step basis, all such numerical trajectories will eventually enter a particular bounded set. This occurs for sufficiently small tolerances independent of the initial conditions. Second, for pairs of arbitrary orders $(p-1,p)$, operating under either error-per-step or error-per-unit-step control, similar results are obtained when an additional structural assumption (that should be valid for many cases of interest) is imposed on the dissipative vector field. Numerical results support both the analysis and the assumptions made.

[1]  Andrew M. Stuart,et al.  Convergence Proofs for Numerical IVP Software , 2000 .

[2]  J. Butcher.  Numerical methods for ordinary differential equations , 2003 .

[3]  Andrew M. Stuart,et al.  Convergence results for the MATLAB ode23 routine , 1998 .

[4]  J. Lambert.  Numerical Methods for Ordinary Differential Equations , 1991 .

[5]  F. Krogh,et al.  Solving Ordinary Differential Equations , 2019, Programming for Computations - Python.

[6]  D. F. Griffiths.  The dynamics of some linear multistep methods with stepsize control , 1989 .

[7]  Andrew M. Stuart.  Probabilistic and deterministic convergence proofs for software for initial value problems , 2004, Numerical Algorithms.

[8]  Ernst Hairer,et al.  Solving Ordinary Differential Equations I: Nonstiff Problems , 2009 .

[9]  Harbir Lamba.  Dynamical Systems and Adaptive Timestepping in ODE Solvers , 2000 .

[10]  Andrew M. Stuart,et al.  Runge-Kutta methods for dissipative and gradient dynamical systems , 1994 .

[11]  Andrew M. Stuart,et al.  The essential stability of local error control for dynamical systems , 1995 .

[12]  R. Lamba.  Convergence Results for the MATLAB ode 23 , 1998 .

[13]  Andrew M. Stuart,et al.  Model Problems in Numerical Stability Theory for Initial Value Problems , 1994, SIAM Rev..

[14]  A. Bountis.  Dynamical Systems And Numerical Analysis , 1997, IEEE Computational Science and Engineering.

[15]  J. Verwer,et al.  Stability of Runge-Kutta Methods for Stiff Nonlinear Differential Equations , 1984 .

[16]  D. Broomhead,et al.  The dynamics of numerics and the numerics of dynamics , 1992 .

[17]  John C. Butcher,et al.  A stability property of implicit Runge-Kutta methods , 1975 .

[18]  Desmond J. Higham,et al.  Analysis of the dynamics of local error control via a piecewise continuous residual , 1998 .