How to Estimate with a Computer program a 3rd order diff eq

Philosophaie · Apr 13, 2009

How to Estimate with a Computer program a 3rd order differential equation with 4 variables and 4 equations? The equations are derived from the geodesics equation and the Schwarzchild metric. The variables are r, theta, phi and t. Any hints would be appreciated.

matematikawan · Apr 14, 2009

Probably if you can write down your equations it will be much clearer what your problem actually is.

Philosophaie · Apr 14, 2009

Equation 1:
[tex]\frac{d}{d\tau}(\frac{2r}{r-2m}(\frac{dr}{d\tau})) + \frac{2m}{(r-2m)^{2}}(\frac{dr}{d\tau})^{2} - r(\frac{d\theta}{d\tau})^{2} - rsin^{2}\theta (\frac{d\phi}{d\tau})^{2} + \frac{mc^{2}}{r^{2}}(\frac{dt}{d\tau})^{2} = 0[/tex]

Equation 2:
[tex]\frac{d}{d\tau}(r^{2}\frac{d\theta}{d\tau}) - r^{2}sin\theta cos\theta (\frac{d\phi}{d\tau}) = 0[/tex]

Equation 3:
[tex]\frac{d}{d\tau}(r^{2}sin^{2}\theta (\frac{d\phi}{d\tau})) = 0[/tex]

Equation 4:
[tex]\frac{d}{d\tau}(\frac{r-2m}{r}(\frac{dt}{d\tau})) = 0[/tex]

where m and c are constants

matematikawan · Apr 14, 2009

It look like your equations have analytical solutions (probably i can be wrong). Start solving equation 4 first, then equation 3.
If I'm not mistaken, for the Schwarzchild metric, there is analytic solution.

But if you still insist on numerical solution you can try Rungge-Kutta fourth order. One of the easiest to program. But you need to rewrite your equations as system of first order DE of the form.

[tex]X'(\tau) = A(\tau) X(\tau)[/tex]

where X is a column vector and A is a matrix.

By the way why do you say your equation is 3rd order?

Philosophaie · Apr 14, 2009

Yes the equation is 2nd order not 3rd.

Philosophaie · Apr 15, 2009

Just checking my math, in the first part of Equation 2:

[tex]\frac{d}{d\tau}(r^{2}\frac{d\theta}{d\tau}) = 2r\frac{dr}{d\tau}\frac{d\theta}{d\tau} + r^{2}(\frac{d\theta}{d\tau})^{2}[/tex]

Is this correct?

matematikawan · Apr 15, 2009

I think from equation 2 and 3 you can obtained much simplier equation, I think.

[tex]\frac{d}{d\tau}(r^{2}\frac{d\theta}{d\tau}) = A \cot (\theta) [/tex]

where A is the arbitrary constant from eq. 2

Philosophaie · Apr 16, 2009

Any suggestions on how to solve for r and [tex]\theta[/tex] from the first two equations if the solution from equation 3 is :

[tex]\frac{d\phi}{d\tau} = \frac{constant3}{r^{2}sin^{2}\theta}[/tex]

and equation 4 is:

[tex]\frac{dt}{d\tau} = \frac{(constant4)r}{r-2m}[/tex]

coomast · Apr 16, 2009

I was reading this post and find it an intresting questing although general relativity is not my field :-) Finding an analytic solution seems to be a challenge and that's what I would like to try. Now it seems that there are 5 variables involved, [itex]\tau[/itex], [itex]r[/itex], [itex]\theta[/itex], [itex]\phi[/itex] and [itex]t[/itex]. Can someone tell me which variable is depending on what?

matematikawan · Apr 19, 2009

Look like tau is the independent variable to me. But what confuse me is that both the letters tau and r look similar in latex.

Philosophaie · Apr 19, 2009

tau takes the place of s in geodesic equations and is called proper time in a "local free-fall frame replacing s by:

[tex]ds^{2} = -c^{2}d\tau^{2}[/tex]