On implicit extrapolation methods for ordinary differential equations

From MaRDI portal





The author considers the numerical integration of the initial value problem for a system of first-order ordinary differential equations using Runge-Kutta methods with subsequent extrapolation. For implicit Runge-Kutta methods the minimum number of iterations for solving the resulting nonlinear algebraic equations without destroying the asymptotic accuracy of the extrapolation process is studied in detail. Implicit Runge-Kutta methods are important in the context of extrapolation because the global error may exhibit an asymptotic expansion in powers of \(h^2\) of the step-size \(h\) which is impossible for explicit methods. For the iteration the simple iteration method, the Newton and a modified Newton method are considered. Finally, the author studies the effect of these iteration methods on the leading error term of the extrapolated solution if only one iteration step is performed. Several numerical examples are given. It would have been interesting to compare the effectivity of the proposed methods with the well-known semi-implicit midpoint rule or others.











This page was built for publication: On implicit extrapolation methods for ordinary differential equations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4545009)