A Runge-Kutta method with stepsize control for separated systems of first-order ODEs (Q1316158)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A Runge-Kutta method with stepsize control for separated systems of first-order ODEs |
scientific article |
Statements
A Runge-Kutta method with stepsize control for separated systems of first-order ODEs (English)
0 references
10 April 1994
0 references
The author considers a method to solve a system of first-order ordinary differential equations (ODEs) separated into two subsystems (1) \({\mathbf y}'= {\mathbf f}(t,{\mathbf y},{\mathbf z})\), \({\mathbf z}'= {\mathbf g}(t,{\mathbf y},{\mathbf z})\); \({\mathbf y}(t_ 0)= {\mathbf y}_ 0\), \({\mathbf z}(t_ 0)= {\mathbf z}_ 0\). First this initial value problem is replaced by the following one on each subinterval \([t_{i-1},t_ i]\): starting with \({\mathbf y}(t_{i-1})= {\mathbf y}_{i-1}\) one integrates the system (2) \({\mathbf y}'= {\mathbf f}(t,{\mathbf y},{\mathbf z}^*(t,{\mathbf y}))\) from \(t= t_{i-1}\) to \(t= t_ i\), where \({\mathbf z}^*(t,{\mathbf y})\) is obtained by integrating the system (3) \({d{\mathbf z}\over d\tau}= {\mathbf g}(\tau,{\mathbf h}(t,\tau,{\mathbf y}),{\mathbf z})\) starting with \({\mathbf z}= {\mathbf z}_{i-1}\) from \(\tau= t_{i-1}\) to \(\tau= t\). Here \[ {\mathbf h}(t,\tau,{\mathbf y})= {\mathbf y}_{i-1}+ \Delta\tau{\mathbf y}_{i- 1}'+\left({\Delta\tau\over \Delta t}\right)^ 2({\mathbf y}-{\mathbf y}_{i-1}- \Delta t{\mathbf y}_{i-1}'), \] where \(\Delta t= t- t_{i- 1}\), \(\Delta\tau= \tau- t_{i-1}\) and \({\mathbf y}_{i-1}'= {\mathbf f}(t,{\mathbf y}_{i-1},{\mathbf z}_{i-1})\). The computed solutions at \(t= t_ i\) are denoted by \({\mathbf y}_ i\) and \({\mathbf z}_ i\), and \({\mathbf z}_ i\) is obtained by integrating system (3) from \(\tau= t_{i-1}\) to \(\tau= t_ i\) using \({\mathbf y}= {\mathbf y}_ i\). Secondly the system (2) is numerically integrated by employing a third- order Runge-Kutta method embedded in a fourth-order Runge-Kutta method. System (3) is integrated by third-order Runge-Kutta formulas. The developed method works with automatic stepsize control, taking both truncation and separation errors into account. The advantage of separating (1) into systems (2) and (3) is that it may be possible to place all of the fast-response equations into (3) and to integrate (2) with a larger stepsize than would have been required to integrate system (1). Moreover, it may be possible to integrate (3) in closed form or with a special method designed for stiff ODEs.
0 references
separation into subsystems
0 references
Runge-Kutta method
0 references
system
0 references
automatic stepsize control
0 references
0 references