I have a very complicated stiff system, which can be expressed as:(adsbygoogle = window.adsbygoogle || []).push({});

[itex]x^{\prime}(t)=A(x,y)-B(x,y)[/itex]

[itex]y^{\prime}(t)=C(x,y)-D(x,y)[/itex]

I decided to solve it with the fixed-point iteration method (http://en.wikipedia.org/wiki/Fixed_point_iteration) but I also have to use adaptive time stepping to iterate the equations. I have decided to chose the time step in the following way:

[itex]dt=\alpha min(\left|\frac{A(x,y)}{x}\right|^{-1},\left|\frac{B(x,y)}{x}\right|^{-1},\left|\frac{C(x,y)}{y}\right|^{-1},\left|\frac{D(x,y)}{y}\right|^{-1},\left|\frac{A(x,y)-B(t,x,y)}{x}\right|^{-1},\left|\frac{C(x,y)-D(x,y)}{y}\right|^{-1})[/itex]

with [itex]\alpha[/itex] a constant less than unity.

What do you think of my choice? Do you think there's a better way to define the time step?

Also, I keep [itex]\alpha[/itex] constant but I noticed I have to decresead for some particular values of x and y. What do you think might a way to determin [itex]\alpha[/itex] as a function of x and y?

**Physics Forums - The Fusion of Science and Community**

The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

# Numerical stability of stiff problem

Can you offer guidance or do you also need help?

Draft saved
Draft deleted

Loading...

Similar Threads - Numerical stability stiff | Date |
---|---|

I Neural Networks vs Traditional Numerical Methods | Feb 7, 2018 |

A Numerical solution techniques for 6th order PDE's | Feb 3, 2018 |

Relation between energy conservation and numerical stability | Nov 10, 2015 |

How to improve stabilities of numerical solutions of PDEs | Jan 27, 2015 |

Numerical Stability of PDE | Aug 22, 2013 |

**Physics Forums - The Fusion of Science and Community**